US Traffic Accidents Severity Analysis & Prediction

Group 7

• Dataset

Table of Content

1. Data Overview
   • 1. Load Data
   • 2. Data Type
   • 3. Statistical View
   • 4. Data Categorizing
2. Data Preprocessing
   • 1. Drop irrelevant columns
   • 2. Drop the column with Missing Value(>40%)
   • 3. Data type correcting
   • 4. Data Imputation
     • 1. Drop all NaN/NA/null
     • 2. Median imputation
     • 3. Mean imputation
3. Data Analysis
   • 1. Basic Analysis
     • 1. What is the distribution of accidents severity?
     • 2. What is the relationship between distance and severity?
   • 2. Location Analysis
     • 1. What are the top 10 states with the most accidents?
     • 3. What are the top 10 cities with the most accidents?
     • 6. What are the accidents distribution by street Side?
   • 3. Time Analysis
     • 1. Accidents yearly change
     • 2. Accidents monthly change
     • 3. Accidents weekly change
     • 4. Accidents hourly change
     • 5. Duration and Severity
   • 4. Environment Analysis
     • 2. Top 15 Wind_Direction
     • 3. Environment Attribute(numerical) distribution and the relationship with Severity
     • 4. Accidents distribution by Sunrise && Sunset
   • 5. Infrastructure Analysis
4. Feature Engineering
   • 1. Feature Choosing
   • 2. One Hot Encoding
   • 3. Label Encoding
5. Modelling
   • 1. Workflow Demonstration
     • 2. Severity Prediction
       • 3. Decision Tree
       • 4. Gradient Boosting Tree
       • 5. Random Forest
       • 6. XGBoost
       • 8. Model Comparison
       • 9. Deal with imbalance data
       • 10. Severity Prediction Visualization
   • 3. Severity Prediction
     • 1. Parameter Turning
     • 2. Model Comparison
6. Conclusion
7. Explainable AI

1. Data Overview

# Import packages

## Basic data processing
import numpy as np
import pandas as pd

## Data Visualization
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline

import plotly.graph_objects as go
import plotly.express as px
from plotly.subplots import make_subplots
#for map
import folium
import folium.plugins

## Modelling
from sklearn.preprocessing import LabelEncoder, OneHotEncoder, StandardScaler
from sklearn.pipeline import make_pipeline
from sklearn.model_selection import train_test_split, ShuffleSplit, cross_val_score, GridSearchCV
from sklearn.metrics import classification_report, confusion_matrix, accuracy_score, f1_score, plot_confusion_matrix, roc_auc_score, balanced_accuracy_score, mean_squared_error, r2_score
from sklearn.tree import DecisionTreeClassifier, DecisionTreeRegressor, export_graphviz, plot_tree
from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor, GradientBoostingClassifier, GradientBoostingRegressor
from xgboost.sklearn import XGBClassifier, XGBRegressor

## Model Explanatory
import eli5
from pdpbox import pdp, get_dataset, info_plots

## Settings
pd.set_option('display.max_columns', 500) # Able to display more columns.

1.1. Load Data

# Load the dataset
data_df = pd.read_csv("US_Accidents_Dec20_Updated.csv")
data_df.info() # show entries, dtypes, memory useage.

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1516064 entries, 0 to 1516063
Data columns (total 47 columns):
 #   Column                 Non-Null Count    Dtype  
---  ------                 --------------    -----  
 0   ID                     1516064 non-null  object 
 1   Severity               1516064 non-null  int64  
 2   Start_Time             1516064 non-null  object 
 3   End_Time               1516064 non-null  object 
 4   Start_Lat              1516064 non-null  float64
 5   Start_Lng              1516064 non-null  float64
 6   End_Lat                1516064 non-null  float64
 7   End_Lng                1516064 non-null  float64
 8   Distance(mi)           1516064 non-null  float64
 9   Description            1516064 non-null  object 
 10  Number                 469969 non-null   float64
 11  Street                 1516064 non-null  object 
 12  Side                   1516064 non-null  object 
 13  City                   1515981 non-null  object 
 14  County                 1516064 non-null  object 
 15  State                  1516064 non-null  object 
 16  Zipcode                1515129 non-null  object 
 17  Country                1516064 non-null  object 
 18  Timezone               1513762 non-null  object 
 19  Airport_Code           1511816 non-null  object 
 20  Weather_Timestamp      1485800 non-null  object 
 21  Temperature(F)         1473031 non-null  float64
 22  Wind_Chill(F)          1066748 non-null  float64
 23  Humidity(%)            1470555 non-null  float64
 24  Pressure(in)           1479790 non-null  float64
 25  Visibility(mi)         1471853 non-null  float64
 26  Wind_Direction         1474206 non-null  object 
 27  Wind_Speed(mph)        1387202 non-null  float64
 28  Precipitation(in)      1005515 non-null  float64
 29  Weather_Condition      1472057 non-null  object 
 30  Amenity                1516064 non-null  bool   
 31  Bump                   1516064 non-null  bool   
 32  Crossing               1516064 non-null  bool   
 33  Give_Way               1516064 non-null  bool   
 34  Junction               1516064 non-null  bool   
 35  No_Exit                1516064 non-null  bool   
 36  Railway                1516064 non-null  bool   
 37  Roundabout             1516064 non-null  bool   
 38  Station                1516064 non-null  bool   
 39  Stop                   1516064 non-null  bool   
 40  Traffic_Calming        1516064 non-null  bool   
 41  Traffic_Signal         1516064 non-null  bool   
 42  Turning_Loop           1516064 non-null  bool   
 43  Sunrise_Sunset         1515981 non-null  object 
 44  Civil_Twilight         1515981 non-null  object 
 45  Nautical_Twilight      1515981 non-null  object 
 46  Astronomical_Twilight  1515981 non-null  object 
dtypes: bool(13), float64(13), int64(1), object(20)
memory usage: 412.1+ MB

The dataset has 1516064 entries and 46 features + 1 target variable. Automatic Type Recognition: 13 of them are bool type, 13 of them are float64, 1 of them is int64, 20 of them are object. Memory usage: 412.1+ MB.

# Have a look
data_df.head(5)

	ID	Severity	Start_Time	End_Time	Start_Lat	Start_Lng	End_Lat	End_Lng	Distance(mi)	Description	Number	Street	Side	City	County	State	Zipcode	Country	Timezone	Airport_Code	Weather_Timestamp	Temperature(F)	Wind_Chill(F)	Humidity(%)	Pressure(in)	Visibility(mi)	Wind_Direction	Wind_Speed(mph)	Precipitation(in)	Weather_Condition	Amenity	Bump	Crossing	Give_Way	Junction	No_Exit	Railway	Roundabout	Station	Stop	Traffic_Calming	Traffic_Signal	Turning_Loop	Sunrise_Sunset	Civil_Twilight	Nautical_Twilight	Astronomical_Twilight
0	A-2716600	3	2016-02-08 00:37:08	2016-02-08 06:37:08	40.10891	-83.09286	40.11206	-83.03187	3.230	Between Sawmill Rd/Exit 20 and OH-315/Olentang...	NaN	Outerbelt E	R	Dublin	Franklin	OH	43017	US	US/Eastern	KOSU	2016-02-08 00:53:00	42.1	36.1	58.0	29.76	10.0	SW	10.4	0.00	Light Rain	False	False	False	False	False	False	False	False	False	False	False	False	False	Night	Night	Night	Night
1	A-2716601	2	2016-02-08 05:56:20	2016-02-08 11:56:20	39.86542	-84.06280	39.86501	-84.04873	0.747	At OH-4/OH-235/Exit 41 - Accident.	NaN	I-70 E	R	Dayton	Montgomery	OH	45424	US	US/Eastern	KFFO	2016-02-08 05:58:00	36.9	NaN	91.0	29.68	10.0	Calm	NaN	0.02	Light Rain	False	False	False	False	False	False	False	False	False	False	False	False	False	Night	Night	Night	Night
2	A-2716602	2	2016-02-08 06:15:39	2016-02-08 12:15:39	39.10266	-84.52468	39.10209	-84.52396	0.055	At I-71/US-50/Exit 1 - Accident.	NaN	I-75 S	R	Cincinnati	Hamilton	OH	45203	US	US/Eastern	KLUK	2016-02-08 05:53:00	36.0	NaN	97.0	29.70	10.0	Calm	NaN	0.02	Overcast	False	False	False	False	True	False	False	False	False	False	False	False	False	Night	Night	Night	Day
3	A-2716603	2	2016-02-08 06:15:39	2016-02-08 12:15:39	39.10148	-84.52341	39.09841	-84.52241	0.219	At I-71/US-50/Exit 1 - Accident.	NaN	US-50 E	R	Cincinnati	Hamilton	OH	45202	US	US/Eastern	KLUK	2016-02-08 05:53:00	36.0	NaN	97.0	29.70	10.0	Calm	NaN	0.02	Overcast	False	False	False	False	True	False	False	False	False	False	False	False	False	Night	Night	Night	Day
4	A-2716604	2	2016-02-08 06:51:45	2016-02-08 12:51:45	41.06213	-81.53784	41.06217	-81.53547	0.123	At Dart Ave/Exit 21 - Accident.	NaN	I-77 N	R	Akron	Summit	OH	44311	US	US/Eastern	KAKR	2016-02-08 06:54:00	39.0	NaN	55.0	29.65	10.0	Calm	NaN	NaN	Overcast	False	False	False	False	False	False	False	False	False	False	False	False	False	Night	Night	Day	Day

1.4. Data Categorizing

• Basic info: 'ID', 'Severity', 'Start_Time', 'End_Time', 'Distance(mi)', 'Description'

• Location(Geographic factor): 'Start_Lat', 'Start_Lng','End_Lat', 'End_Lng', 'Number', 'Street','Side', 'City', 'County', 'State', 'Zipcode', 'Country', 'Timezone', 'Airport_Code'

• Environment(Environmental factor): 'Weather_Timestamp', 'Temperature(F)', 'Wind_Chill(F)', 'Humidity(%)', 'Pressure(in)', 'Visibility(mi)', 'Wind_Direction', 'Wind_Speed(mph)', 'Precipitation(in)', 'Weather_Condition', 'Sunrise_Sunset', 'Civil_Twilight', 'Nautical_Twilight', 'Astronomical_Twilight'

• Infrastructure(Infrastructural factor): 'Amenity', 'Bump', 'Crossing', 'Give_Way', 'Junction', 'No_Exit', 'Railway', 'Roundabout', 'Station', 'Stop', 'Traffic_Calming', 'Traffic_Signal', 'Turning_Loop'

# Category Visualization
# https://plotly.com/python/sunburst-charts/
data_category = dict(
    character=["Basic", "Location", "Environment", "Infrastructure", 'ID', 'Severity', 'Start_Time', 'End_Time', 'Distance(mi)', 'Description', 'Start_Lat', 'Start_Lng','End_Lat', 'End_Lng', 'Number', 'Street','Side', 'City', 'County', 'State', 'Zipcode', 'Country', 'Timezone', 'Airport_Code', 'Weather_Timestamp', 'Temperature(F)', 'Wind_Chill(F)', 'Humidity(%)', 'Pressure(in)', 'Visibility(mi)', 'Wind_Direction', 'Wind_Speed(mph)', 'Precipitation(in)', 'Weather_Condition', 'Sunrise_Sunset', 'Civil_Twilight', 'Nautical_Twilight', 'Astronomical_Twilight', 'Amenity', 'Bump', 'Crossing', 'Give_Way', 'Junction', 'No_Exit', 'Railway', 'Roundabout', 'Station', 'Stop', 'Traffic_Calming', 'Traffic_Signal', 'Turning_Loop'],
    parent=["","","","","Basic", "Basic", "Basic", "Basic", "Basic", "Basic", "Location", "Location", "Location", "Location", "Location", "Location", "Location", "Location", "Location", "Location", "Location", "Location", "Location", "Location", "Environment", "Environment", "Environment", "Environment", "Environment", "Environment", "Environment", "Environment", "Environment", "Environment", "Environment", "Environment", "Environment", "Environment", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure", "Infrastructure"],
)

category_fig =px.sunburst(
    data_category,
    names='character',
    parents='parent'
)
category_fig.update_layout(
    autosize=False,
    width=600,
    height=600,
    title={
        'text': "Data Categorizing",
        'y':0.95,
        'x':0.5,
        'xanchor': 'center',
        'yanchor': 'top'})
category_fig.show()

2. Data Preprocessing

2.1. Drop irrelevant columns

# Irrelevant columns
'''
ID: ID is unique and meaningless for the dataset.
Description: I don't do text mining, therefore It's useless.
Country: All the data is from US, therefore all the data is the same.
Weather_Timestamp: The timestamp of weather observation record. It's useless here.
'''
irrelavant_columns = ['ID','Description','Country','Weather_Timestamp']
data_preprocessed_df = data_df.drop(irrelavant_columns, axis=1)

2.2. Drop the column with Missing Value(>40%)

# Replace the empty data with NaN
data_preprocessed_df.replace("", float("NaN"), inplace=True)
data_preprocessed_df.replace(" ", float("NaN"), inplace=True)

# Count missing value(NaN, na, null, None) of each columns, Then transform the result to a pandas dataframe. 
count_missing_value = data_preprocessed_df.isna().sum() / data_preprocessed_df.shape[0] * 100
count_missing_value_df = pd.DataFrame(count_missing_value.sort_values(ascending=False), columns=['Missing%'])

# Visualize the percentage(>0) of Missing value in each column.
missing_value_df = count_missing_value_df[count_missing_value_df['Missing%'] > 0]

plt.figure(figsize=(15, 10)) # Set the figure size
missing_value_graph = sns.barplot(y = missing_value_df.index, x = "Missing%", data=missing_value_df, orient="h")
missing_value_graph.set_title("Percentage Missing value of each feature", fontsize = 20)
missing_value_graph.set_ylabel("Features")

Text(0, 0.5, 'Features')

## Drop the column with Missing value(>40%)
missing_value_40_df = count_missing_value_df[count_missing_value_df['Missing%'] > 40]
data_preprocessed_df.drop(missing_value_40_df.index, axis=1, inplace=True)
missing_value_40_df

	Missing%
Number	69.000715

2.3. Data type correcting

# Convert Time to datetime64[ns]
data_preprocessed_df['Start_Time'] = pd.to_datetime(data_preprocessed_df['Start_Time'])
data_preprocessed_df['End_Time'] = pd.to_datetime(data_preprocessed_df['End_Time'])

2.4. Data Imputation

Choose the suitable imputation tech which can highly represent the central tendency of the data.

# Display all the missing value
missing_value_df

	Missing%
Number	69.000715
Precipitation(in)	33.675953
Wind_Chill(F)	29.637007
Wind_Speed(mph)	8.499773
Humidity(%)	3.001786
Visibility(mi)	2.916170
Weather_Condition	2.902714
Temperature(F)	2.838469
Wind_Direction	2.760965
Pressure(in)	2.392643
Airport_Code	0.280199
Timezone	0.151841
Zipcode	0.061673
Nautical_Twilight	0.005475
Civil_Twilight	0.005475
Sunrise_Sunset	0.005475
Astronomical_Twilight	0.005475
City	0.005475

# Categorize the missing value to numerical and categorical for imputation purpose
numerical_missing = ['Wind_Speed(mph)', 'End_Lng', 'End_Lat', 'Visibility(mi)','Humidity(%)', 'Temperature(F)', 'Pressure(in)']
categorical_missing = ['Weather_Condition','Wind_Direction', 'Sunrise_Sunset', 'Civil_Twilight', 'Nautical_Twilight', 'Astronomical_Twilight', 'Side']

2.4.1. Drop all NaN/NA/null

# Drop all the instance with NaN/NA/null
data_preprocessed_dropNaN_df = data_preprocessed_df.dropna()
data_preprocessed_dropNaN_df.reset_index(drop=True, inplace=True)

2.4.2. Median imputation

# Imputation by corresponding class Median value 
data_preprocessed_median_df = data_preprocessed_df.copy()

# For numerical columns
for column_name in numerical_missing:
    data_preprocessed_median_df[column_name] = data_preprocessed_median_df.groupby('Severity')[column_name].transform(lambda x:x.fillna(x.median()))

# # For categorical columns(Majority value imputation)
# https://medium.com/analytics-vidhya/best-way-to-impute-categorical-data-using-groupby-mean-mode-2dc5f5d4e12d
for column_name in categorical_missing:
    data_preprocessed_median_df[column_name] = data_preprocessed_median_df.groupby('Severity')[column_name].transform(lambda x:x.fillna(x.fillna(x.mode().iloc[0])))

# Drop NaN and reset index
data_preprocessed_median_df.dropna(inplace=True)

2.4.3. Mean imputation

# Imputation by corresponding class Mean value 
data_preprocessed_mean_df = data_preprocessed_df.copy()

# For numerical columns
for column_name in numerical_missing:
    data_preprocessed_mean_df[column_name] = data_preprocessed_mean_df.groupby('Severity')[column_name].transform(lambda x:x.fillna(x.mean()))

# For categorical columns(Majority value imputation)
for column_name in categorical_missing:
    data_preprocessed_mean_df[column_name] = data_preprocessed_mean_df.groupby('Severity')[column_name].transform(lambda x:x.fillna(x.fillna(x.mode().iloc[0])))
    
# Drop NaN 
data_preprocessed_mean_df.dropna(inplace=True)

After Data Imputation:

• data_preprocessed_dropNaN_df: The dataset that simply drop all the NaN values.
• data_preprocessed_median_df: The dataset that use grouped median imputation for numerical data and grouped majority value imputation for categorical data.
• data_preprocessed_mean_df: The dataset that use grouped mean imputation for numerical data and grouped majority value imputation for categorical data.

# Save these datasets to local
#data_preprocessed_dropNaN_df.to_csv('preprocessed_dropNaN.csv', index=False)
#data_preprocessed_median_df.to_csv('preprocessed_median.csv', index=False)
#data_preprocessed_mean_df.to_csv('preprocessed_mean.csv', index=False)

# Choose the best dataset base on the performance of modeling
data_best_df = data_preprocessed_dropNaN_df.copy()
#data_best_df = data_preprocessed_dropNaN_df[data_preprocessed_dropNaN_df['City'] == 'Orlando'].copy()
#data_best_df = data_preprocessed_median_df[data_preprocessed_dropNaN_df['City'] == 'Orlando'].copy()
#data_best_df = data_preprocessed_mean_df[data_preprocessed_dropNaN_df['City'] == 'Orlando'].copy()

# Reset index
data_best_df.reset_index(inplace=True)

3. Data Analysis

3.1. Basic Analysis

3.1.1. What is the distribution of accidents severity?

# Count the number of each severity, transform the result to pandas dataframe
severity_counts = data_best_df["Severity"].value_counts()
severity_counts_df = pd.DataFrame(severity_counts)

# Calculate the proportion of each Severity
severity_percentage_df = severity_counts_df / sum(severity_counts_df["Severity"]) * 100

# Visualize the distribution of accidents severity
severity_fig = make_subplots(
    rows=1, cols=2, 
    specs=[[{"type": "xy"}, {"type": "domain"}]])

severity_fig.add_trace(go.Bar(x=severity_counts_df.index, 
                              y=severity_counts_df["Severity"],
                              text=severity_counts_df["Severity"],
                              textposition='outside',
                              showlegend=False),
                              1, 1)

severity_fig.add_trace(go.Pie(labels=severity_percentage_df.index, 
                     values=severity_percentage_df["Severity"],
                     showlegend=True),
                     1, 2)

severity_fig.update_layout(
                  height=600, 
                  width=1500,
                  title={
                  'text': "The distribution of accidents severity",
                  'font': {'size': 24},
                  'y':0.95,
                  'x':0.5,
                  'xanchor': 'center',
                  'yanchor': 'top'},
                  xaxis1_title = 'Severity',
                  yaxis1_title = 'Counts',
                  legend_title_text="Severity"
                 )
severity_fig.update_xaxes(type='category')
severity_fig.show()

As we can see from the graph, level 2 is the most frequent severity which is 76.1% of the total. That means our target variable(label) is quite unbalanced.

3.1.2. What is the relationship between distance and severity?

# Calculate the mean distance of each Severity
mean_distance = data_best_df.groupby('Severity')["Distance(mi)"].mean().round(2)
mean_distance_df = pd.DataFrame(mean_distance)

mean_distance_fig = px.bar(mean_distance_df, 
                                   x = mean_distance_df.index, 
                                   y = "Distance(mi)", 
                                   labels={"index": "Severity"},
                                   text="Distance(mi)")
mean_distance_fig.update_layout(
    autosize=False,
    width=600,
    height=500,
    title={
        'text': "Mean Distance(mi) of each Severity",
        'y':0.95,
        'x':0.5,
        'font': {'size': 24},
        'xanchor': 'center',
        'yanchor': 'top'},
    xaxis={
        'type':'category'
    })
mean_distance_fig.show()

Generally speaking, the longer the influence distance the higher the severity level will be.

3.2. Location Analysis

3.2.1. What are the top 10 states with the most accidents?

# Overview of the US traffic accidents 
# https://plotly.com/python/choropleth-maps/
state_accidents_count = data_best_df["State"].value_counts()
fig = go.Figure(data=go.Choropleth(locations=state_accidents_count.index, 
                                           z=state_accidents_count.values.astype(float), 
                                           locationmode="USA-states", 
                                           colorscale="Reds",
                                           colorbar_title = "Frequency"
                                  ))
fig.update_layout(
    height=600, 
    width=1500,
    title={
        'text': "Frequency distribution of US Accidents",
        'y':0.9,
        'x':0.5,
        'font': {'size': 24},
        'xanchor': 'center',
        'yanchor': 'top'},
        geo_scope="usa")
fig.show()

# Top 10 States with the Most Accidents
state_accidents_count_top10 = state_accidents_count[:10]
state_accidents_count_top10_df = pd.DataFrame(state_accidents_count_top10)

state_accidents_count_top10_fig = px.bar(state_accidents_count_top10_df, 
                                         x = state_accidents_count_top10_df.index, 
                                         y = "State", 
                                         labels={"index": "State", "State": "Counts"},
                                         text="State")
state_accidents_count_top10_fig.update_layout(
    autosize=False,
    width=1000,
    height=600,
    title={
        'text': "Top 10 States with the Most Accidents",
        'y':0.95,
        'x':0.5,
        'font': {'size': 24},
        'xanchor': 'center',
        'yanchor': 'top'})
state_accidents_count_top10_fig.update_yaxes(categoryorder = "total ascending")
state_accidents_count_top10_fig.show()

# Top 10 States with the Most Accidents in a view of severity 
plt.figure(figsize=(20, 8))
ax = sns.countplot(x="State", 
              data=data_best_df, 
              order=data_best_df['State'].value_counts()[:10].index, 
              hue='Severity',
              palette='tab10')
plt.title("Top 10 States with the Most Accidents", fontsize = 22)
#for p in ax.patches:
#        ax.annotate(p.get_height(), (p.get_x(), p.get_height()+1000))
plt.show()

3.2.3. What are the top 10 cities with the most accidents?

# Top 10 Cities with the Most Accidents
city_accidents_count = data_best_df["City"].value_counts()
city_accidents_count_top10 = city_accidents_count[:10]
city_accidents_count_top10_df = pd.DataFrame(city_accidents_count_top10)

city_accidents_count_top10_fig = px.bar(city_accidents_count_top10_df, 
                                        x = city_accidents_count_top10_df.index, 
                                        y = "City", 
                                        labels={"index": "City", "City": "Counts"},
                                        text="City")
city_accidents_count_top10_fig.update_layout(
    autosize=False,
    width=1000,
    height=600,
    title={
        'text': "Top 10 Cities with the Most Accidents",
        'y':0.95,
        'x':0.5,
        'font': {'size': 24},
        'xanchor': 'center',
        'yanchor': 'top'})
city_accidents_count_top10_fig.update_yaxes(categoryorder = "total ascending")
city_accidents_count_top10_fig.show()

# Top 10 Cities with the Most Accidents in a view of severity
plt.figure(figsize=(20, 8))
ax = sns.countplot(x="City", 
                   data = data_best_df, 
                   order=data_best_df['City'].value_counts()[:10].index, 
                   hue = 'Severity')
plt.title("Top 10 Cities with the Most Accidents", fontsize = 22)
for p in ax.patches:
        ax.annotate(p.get_height(), (p.get_x(), p.get_height()+200))
plt.show()

3.2.6. What are the accidents distribution by street Side?

# Accidents distribution by street Side
# Set up the matplotlib figure
f, axes = plt.subplots(nrows=1, ncols=2, figsize=(20, 8))
# Pie chart
data_best_df["Side"].value_counts().plot.pie(autopct="%.1f%%", ylabel='', ax=axes[0])

sns.countplot(x="Side", 
              data = data_best_df, 
              order=data_best_df['Side'].value_counts().index, 
              hue = 'Severity',
             ax=axes[1])
for p in axes[1].patches:
        axes[1].annotate(p.get_height(), (p.get_x()+0.05, p.get_height()+100))

# Common title
plt.suptitle("Accidents distribution by street Side", y=0.95, fontsize=20)
plt.show()

Most of the accidents happened at the right side of the road which is quite an interesting finding.

3.3. Time Analysis

3.3.4. Hour

# Extract Hour (Weekday)
weekdays_lst = [ 'Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday']
hour_accidents_weekdays_count = data_best_df[data_best_df['Week'].isin(weekdays_lst)]["Start_Time"].dt.hour
hour_accidents_weekdays_count_df = pd.DataFrame(hour_accidents_weekdays_count.value_counts())
hour_accidents_weekdays_count_df.sort_index(inplace=True)

hour_accidents_weekdays_count_fig = px.bar(hour_accidents_weekdays_count_df, 
                                   x = hour_accidents_weekdays_count_df.index, 
                                   y = "Start_Time", 
                                   labels={"index": "Hour", "Start_Time": "Counts"},
                                   text="Start_Time")
hour_accidents_weekdays_count_fig.update_layout(
    autosize=False,
    width=1000,
    height=500,
    title={
        'text': "Accidents hourly change(Weekdays)",
        'y':0.95,
        'x':0.5,
        'font': {'size': 24},
        'xanchor': 'center',
        'yanchor': 'top'},
    xaxis={
        'type':'category'
    })
hour_accidents_weekdays_count_fig.show()

It seems that 7:00/8:00(Go to work) and 16:00/17:00(go back home) are the time where accidents happen during a day.

# Extract Hour (Weekend)
weekend_lst = ['Saturday', 'Sunday']
hour_accidents_weekend_count = data_best_df[data_best_df['Week'].isin(weekend_lst)]["Start_Time"].dt.hour
hour_accidents_weekend_count_df = pd.DataFrame(hour_accidents_weekend_count.value_counts())
hour_accidents_weekend_count_df.sort_index(inplace=True)

hour_accidents_weekend_count_fig = px.bar(hour_accidents_weekend_count_df, 
                                   x = hour_accidents_weekend_count_df.index, 
                                   y = "Start_Time", 
                                   labels={"index": "Hour", "Start_Time": "Counts"},
                                   text="Start_Time")
hour_accidents_weekend_count_fig.update_layout(
    autosize=False,
    width=1000,
    height=500,
    title={
        'text': "Accidents hourly change(Weekend)",
        'y':0.95,
        'x':0.5,
        'font': {'size': 24},
        'xanchor': 'center',
        'yanchor': 'top'},
    xaxis={
        'type':'category'
    })
hour_accidents_weekend_count_fig.show()

The peak of the weekend is during 12:00 ~ 16:00.

# Weekly view with hours
data_best_df.groupby('Week')['Hour'].value_counts().unstack().reindex(week_order).plot.bar(
    figsize=(22, 8),
    ylabel='Counts',
    width=.9
)
plt.title("Accidents Weekly change in a view of hour", fontsize = 22)
plt.show()

The decrease of the accidents number at weekends mainly because the decrease at 7:00/8:00(Go to work) and 16:00/17:00(go back home). The peak of the weekend is during 12:00 ~ 16:00.

3.4. Environment Analysis

# Weather condition by mean of the Severity
weather_mean_severity = data_best_df.groupby('Weather_Condition')['Severity'].mean().sort_values(ascending=False)
weather_mean_severity_df = pd.DataFrame(weather_mean_severity[:25])

plt.figure(figsize=(15, 10)) # Set the figure size
weather_mean_severity_graph = sns.barplot(y = weather_mean_severity_df.index, x = "Severity", data=weather_mean_severity_df, orient="h")
weather_mean_severity_graph.set_title("Weather Condition with mean of the severity", fontsize = 20)
weather_mean_severity_graph.set_ylabel("Weather_Condition")

Text(0, 0.5, 'Weather_Condition')

Freezing rain with windy, light blowing Snow and patches of fog with windy are the top 3 dangeous weather condition.

Temperature(F): lower temperature -> higher Severity
Humidity(%): higher humidity -> higher Severity
Pressure(in): lower pressure -> higher Severity
Visibility(mi): lower visibility -> higher Severity
Wind_Speed(mph): higher wind speed -> higher Severity
In summary, all the results fit with cold, chill, freezing weather condition. Eg. Freezing rain with windy, snow and so on.

3.4.4. Accidents distribution by Sunrise && Sunset

# Accidents distribution by Sunrise && Sunset
# Set up the matplotlib figure
f, axes = plt.subplots(nrows=1, ncols=2, figsize=(20, 8))
# Pie chart
data_best_df["Sunrise_Sunset"].value_counts().plot.pie(autopct="%.1f%%", ylabel='', ax=axes[0])

sns.countplot(x="Sunrise_Sunset", 
              data = data_best_df, 
              order=data_best_df['Sunrise_Sunset'].value_counts().index, 
              hue = 'Severity',
             ax=axes[1])
for p in axes[1].patches:
        axes[1].annotate(p.get_height(), (p.get_x()+0.025, p.get_height()+100))

# Common title
plt.suptitle("Accidents distribution by Sunrise && Sunset", y=0.95, fontsize=20)
plt.show()

Accidents mostly happened at daytime.

4. Feature Engineering

## Only choose a city because of the resources limitation.
data_best_df = data_preprocessed_dropNaN_df[data_preprocessed_dropNaN_df['City'] == 'Charlotte'].copy()
#data_best_df = data_preprocessed_median_df[data_preprocessed_dropNaN_df['City'] == 'Orlando'].copy()
#data_best_df = data_preprocessed_mean_df[data_preprocessed_dropNaN_df['City'] == 'Orlando'].copy()

# Reset index
data_best_df.reset_index(inplace=True)

4.1. Feature choosing

# Choose relevant features
relevant_features = ['Severity', 'Start_Time', 'End_Time', 'Start_Lat', 'Start_Lng','Side',
       'Temperature(F)', 'Humidity(%)', 'Pressure(in)', 'Visibility(mi)',
       'Wind_Direction', 'Wind_Speed(mph)', 'Weather_Condition', 'Amenity',
       'Bump', 'Crossing', 'Give_Way', 'Junction', 'No_Exit', 'Railway',
       'Roundabout', 'Station', 'Stop', 'Traffic_Calming', 'Traffic_Signal',
       'Turning_Loop', 'Sunrise_Sunset']
data_modelling_df = data_best_df[relevant_features].copy()

# Duration = End_Time - Start_Time; Create a new feature for modeling.
data_modelling_df['Duration'] = (data_modelling_df['End_Time'] - data_modelling_df['Start_Time']).dt.total_seconds() / 3600
data_modelling_df.drop('End_Time', axis=1, inplace=True)

# Transform Month/week/Hour to different features
data_modelling_df["Month"] = data_modelling_df["Start_Time"].dt.month
data_modelling_df["Week"] = data_modelling_df["Start_Time"].dt.dayofweek
data_modelling_df["Hour"] = data_modelling_df["Start_Time"].dt.hour
data_modelling_df.drop("Start_Time", axis=1, inplace=True)

4.2. One Hot Encoding

# Select features that are suitable for One Hot Encoding
one_hot_features = ['Wind_Direction', 'Weather_Condition']

# Wind_Direction Categorizing
data_modelling_df.loc[data_modelling_df['Wind_Direction'].str.startswith('C'), 'Wind_Direction'] = 'C' #Calm
data_modelling_df.loc[data_modelling_df['Wind_Direction'].str.startswith('E'), 'Wind_Direction'] = 'E' #East, ESE, ENE
data_modelling_df.loc[data_modelling_df['Wind_Direction'].str.startswith('W'), 'Wind_Direction'] = 'W' #West, WSW, WNW
data_modelling_df.loc[data_modelling_df['Wind_Direction'].str.startswith('S'), 'Wind_Direction'] = 'S' #South, SSW, SSE
data_modelling_df.loc[data_modelling_df['Wind_Direction'].str.startswith('N'), 'Wind_Direction'] = 'N' #North, NNW, NNE
data_modelling_df.loc[data_modelling_df['Wind_Direction'].str.startswith('V'), 'Wind_Direction'] = 'V' #Variable

# Weather_Condition Categorizing
# Fair, Cloudy, Clear, Overcast, Snow, Haze, Rain, Thunderstorm, Windy, Hail, Thunder, Dust, Tornado
data_modelling_df['Weather_Fair'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Fair', case=False, na = False), 1, 0)
data_modelling_df['Weather_Cloudy'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Cloudy', case=False, na = False), 1, 0)
data_modelling_df['Weather_Clear'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Clear', case=False, na = False), 1, 0)
data_modelling_df['Weather_Overcast'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Overcast', case=False, na = False), 1, 0)
data_modelling_df['Weather_Snow'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Snow|Wintry|Sleet', case=False, na = False), 1, 0)
data_modelling_df['Weather_Haze'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Smoke|Fog|Mist|Haze', case=False, na = False), 1, 0)
data_modelling_df['Weather_Rain'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Rain|Drizzle|Showers', case=False, na = False), 1, 0)
data_modelling_df['Weather_Thunderstorm'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Thunderstorms|T-Storm', case=False, na = False), 1, 0)
data_modelling_df['Weather_Windy'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Windy|Squalls', case=False, na = False), 1, 0)
data_modelling_df['Weather_Hail'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Hail|Ice Pellets', case=False, na = False), 1, 0)
data_modelling_df['Weather_Thunder'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Thunder', case=False, na = False), 1, 0)
data_modelling_df['Weather_Dust'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Dust', case=False, na = False), 1, 0)
data_modelling_df['Weather_Tornado'] = np.where(data_modelling_df['Weather_Condition'].str.contains('Tornado', case=False, na = False), 1, 0)

# Transform the one-hot features, then delete them
onehot_df = pd.get_dummies(data_modelling_df['Wind_Direction'], prefix='Wind')
data_modelling_df = pd.concat([data_modelling_df, onehot_df], axis=1)
data_modelling_df.drop(one_hot_features, axis=1, inplace=True)

4.3. Label Encoding

# Select features that are suitable for Label Encoding
label_encoding_features = ['Side', 'Amenity','Bump', 'Crossing', 'Give_Way', 'Junction', 'No_Exit', 'Railway','Roundabout', 'Station', 'Stop', 'Traffic_Calming', 'Traffic_Signal','Turning_Loop', 'Sunrise_Sunset']

# Label Encoding
for feature in label_encoding_features:
    data_modelling_df[feature] = LabelEncoder().fit_transform(data_modelling_df[feature])

data_modelling_df

	Severity	Start_Lat	Start_Lng	Side	Temperature(F)	Humidity(%)	Pressure(in)	Visibility(mi)	Wind_Speed(mph)	Amenity	Bump	Crossing	Give_Way	Junction	No_Exit	Railway	Roundabout	Station	Stop	Traffic_Calming	Traffic_Signal	Turning_Loop	Sunrise_Sunset	Duration	Month	Week	Hour	Weather_Fair	Weather_Cloudy	Weather_Clear	Weather_Overcast	Weather_Snow	Weather_Haze	Weather_Rain	Weather_Thunderstorm	Weather_Windy	Weather_Hail	Weather_Thunder	Weather_Dust	Weather_Tornado	Wind_C	Wind_E	Wind_N	Wind_S	Wind_V	Wind_W
0	2	35.311788	-80.969496	1	45.0	90.0	29.89	5.00	11.5	0	0	0	0	0	0	0	0	1	0	0	0	0	1	6.000000	12	1	6	0	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0
1	2	35.154630	-80.794820	1	45.0	90.0	29.89	5.00	11.5	0	0	1	0	0	0	0	0	0	0	0	1	0	1	6.000000	12	1	6	0	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0
2	2	35.154510	-80.794920	1	45.0	90.0	29.89	5.00	11.5	0	0	1	0	0	0	0	0	0	0	0	1	0	1	6.000000	12	1	6	0	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0
3	2	35.311506	-80.970119	1	45.0	90.0	29.89	5.00	11.5	0	0	0	0	0	0	0	0	1	0	0	0	0	1	6.000000	12	1	6	0	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0
4	2	35.202560	-80.726070	0	45.0	90.0	29.89	5.00	11.5	0	0	1	0	0	0	0	0	0	0	0	1	0	1	6.000000	12	1	6	0	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
10307	2	35.195320	-80.883840	1	90.0	52.0	29.18	10.00	15.0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.486389	8	3	16	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	1	0	0
10308	2	35.262040	-80.967620	1	73.0	87.0	29.21	0.75	3.0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0.491944	8	3	17	0	0	0	0	0	0	1	0	0	0	1	0	0	0	0	1	0	0	0
10309	4	35.184900	-80.888080	1	77.0	82.0	29.21	10.00	0.0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	4.987222	8	3	18	0	1	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0
10310	3	35.264430	-80.842970	1	74.0	87.0	29.24	10.00	6.0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.483333	8	4	22	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0
10311	3	35.271640	-80.851470	1	73.0	87.0	29.25	0.50	9.0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0.500000	8	4	19	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	1	0	0

10312 rows × 46 columns

5. Modelling

5.1. Workflow Demonstration

5.1.2. Severity Prediction

# Train/Test Split
X_cla = data_modelling_df.drop("Severity", axis=1)
Y_cla = data_modelling_df.Severity
x_train_cla, x_test_cla, y_train_cla, y_test_cla = train_test_split(X_cla, Y_cla, test_size = 0.4, random_state=0, stratify=Y_cla)
print(f'Train Cla: {x_train_cla.shape} \n Test Cla: {x_test_cla.shape}')
cla_feature_names = x_train_cla.columns.tolist()

Train Cla: (6187, 45) 
 Test Cla: (4125, 45)

# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.plot_confusion_matrix.html#sklearn.metrics.plot_confusion_matrix
# normalize must be one of {'true', 'pred', 'all', None}
'''
Description: Plot the confusion matrix
Args:
    classifier: The classifier
Return: None
'''
def draw_confusion_matrix(classifier):
    fig, ax = plt.subplots(figsize=(12, 6))
    plot_confusion_matrix(classifier, x_test_cla, y_test_cla, cmap=plt.cm.Blues, normalize=None, ax=ax)
    ax.set_title("Confusion Matrix", fontsize = 15)
    plt.show()

5.1.2.3. Decision Tree

# Decision Tree Classification
dt_cla = DecisionTreeClassifier(random_state=0)
dt_cla.fit(x_train_cla, y_train_cla)
dt_cla_predictions = dt_cla.predict(x_test_cla)
dt_cla_results = classification_report(y_test_cla, dt_cla_predictions, zero_division=True, output_dict=True)

# Confusion matrix and Classification report
draw_confusion_matrix(dt_cla)
print(classification_report(y_test_cla, dt_cla_predictions, zero_division=True))

# balanced_accuracy
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.balanced_accuracy_score.html#sklearn.metrics.balanced_accuracy_score
dt_cla_balanced_accuracy = balanced_accuracy_score(y_test_cla, dt_cla_predictions)
print(f'balanced_accuracy: {dt_cla_balanced_accuracy}')

# ROC_AUC score
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html#sklearn.metrics.roc_auc_score
dt_cla_roc_ovo_macro = roc_auc_score(y_test_cla, dt_cla.predict_proba(x_test_cla), multi_class='ovo', average='macro') #Insensitive to class imbalance when average == 'macro'
dt_cla_roc_ovr_weighted = roc_auc_score(y_test_cla, dt_cla.predict_proba(x_test_cla), multi_class='ovr', average='weighted') #Sensitive to class imbalance even when average == 'macro'
print(f"roc_ovo_macro: {dt_cla_roc_ovo_macro}") 
print(f"roc_ovr_weighted: {dt_cla_roc_ovr_weighted}")

# Show feature importance as a table
eli5.show_weights(dt_cla, feature_names = cla_feature_names)

              precision    recall  f1-score   support

           1       0.55      0.55      0.55       188
           2       0.93      0.93      0.93      3608
           3       0.35      0.36      0.35       279
           4       0.29      0.32      0.30        50

    accuracy                           0.87      4125
   macro avg       0.53      0.54      0.54      4125
weighted avg       0.87      0.87      0.87      4125

balanced_accuracy: 0.5404424097682046
roc_ovo_macro: 0.693470558724398
roc_ovr_weighted: 0.7295847475566548

Weight	Feature
0.1513	Duration
0.1114	Start_Lat
0.1003	Month
0.0989	Start_Lng
0.0846	Hour
0.0768	Pressure(in)
0.0703	Humidity(%)
0.0588	Temperature(F)
0.0472	Week
0.0459	Wind_Speed(mph)
0.0414	Traffic_Signal
0.0184	Visibility(mi)
0.0122	Side
0.0119	Wind_W
0.0107	Weather_Cloudy
0.0096	Wind_S
0.0082	Wind_N
0.0062	Weather_Rain
0.0059	Junction
0.0051	Wind_E
… 25 more …

Month <= 8.500  (55.1%)
    Duration <= 1.000  (50.2%)
        Week <= 4.500  (41.0%)
            Traffic_Signal <= 0.500  (17.1%)
                Side <= 0.500  (3.3%)
                    Start_Lat <= 35.375  (3.2%)
                        Month <= 7.500  (3.2%)
                            Wind_V <= 0.500  (3.1%)
                                Pressure(in) <= 29.485  (3.0%)
                                    Hour <= 21.500  (2.9%)
                                        Temperature(F) <= 88.500  (2.8%)
                                            Week <= 2.500  (1.8%)
                                                Duration <= 0.494  (0.6%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Duration > 0.494  (1.2%)
                                                    Duration <= 0.496  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Duration > 0.496  (1.2%)
                                                        Start_Lat <= 35.109  (0.0%)
                                                            Humidity(%) <= 80.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Humidity(%) > 80.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lat > 35.109  (1.1%)
                                                            Start_Lng <= -80.795  (0.6%)
                                                                Start_Lng <= -80.801  (0.5%)
                                                                    Hour <= 19.500  (0.5%)
                                                                        Temperature(F) <= 57.500  (0.0%)
                                                                            Hour <= 12.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                            Hour > 12.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Temperature(F) > 57.500  (0.5%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Hour > 19.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Start_Lng > -80.801  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Start_Lng > -80.795  (0.6%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Week > 2.500  (1.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Temperature(F) > 88.500  (0.1%)
                                            Temperature(F) <= 89.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Temperature(F) > 89.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Hour > 21.500  (0.0%)
                                        Pressure(in) <= 29.345  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Pressure(in) > 29.345  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Pressure(in) > 29.485  (0.1%)
                                    Hour <= 15.000  (0.1%)
                                        Start_Lng <= -80.872  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                        Start_Lng > -80.872  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Hour > 15.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                            Wind_V > 0.500  (0.1%)
                                Month <= 4.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Month > 4.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                        Month > 7.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                    Start_Lat > 35.375  (0.1%)
                        Wind_N <= 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                        Wind_N > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                Side > 0.500  (13.7%)
                    Temperature(F) <= 68.500  (6.5%)
                        Pressure(in) <= 30.000  (6.4%)
                            Start_Lat <= 35.065  (0.1%)
                                Pressure(in) <= 29.220  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Pressure(in) > 29.220  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Start_Lat > 35.065  (6.3%)
                                Humidity(%) <= 21.000  (0.1%)
                                    Start_Lng <= -80.826  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Start_Lng > -80.826  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Humidity(%) > 21.000  (6.3%)
                                    Hour <= 1.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Hour > 1.000  (6.2%)
                                        Junction <= 0.500  (4.9%)
                                            Start_Lat <= 35.278  (2.9%)
                                                Wind_Speed(mph) <= 14.500  (2.7%)
                                                    Wind_Speed(mph) <= 1.500  (0.3%)
                                                        Week <= 1.500  (0.1%)
                                                            Hour <= 6.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Hour > 6.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Week > 1.500  (0.2%)
                                                            Temperature(F) <= 62.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Temperature(F) > 62.500  (0.1%)
                                                                Start_Lat <= 35.227  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lat > 35.227  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Wind_Speed(mph) > 1.500  (2.4%)
                                                        Humidity(%) <= 82.500  (1.3%)
                                                            Hour <= 18.500  (1.2%)
                                                                Start_Lng <= -80.970  (0.0%)
                                                                    Pressure(in) <= 29.315  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Pressure(in) > 29.315  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                Start_Lng > -80.970  (1.2%)
                                                                    Humidity(%) <= 26.500  (0.0%)
                                                                        Temperature(F) <= 66.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Temperature(F) > 66.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                    Humidity(%) > 26.500  (1.1%)
                                                                        Start_Lat <= 35.147  (0.1%)
                                                                            Humidity(%) <= 57.000  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                            Humidity(%) > 57.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Start_Lat > 35.147  (1.0%)
                                                                            Hour <= 17.500  (0.9%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                            Hour > 17.500  (0.1%)
                                                                                Start_Lng <= -80.839  (0.0%)
                                                                                    Month <= 3.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                    Month > 3.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                                Start_Lng > -80.839  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Hour > 18.500  (0.1%)
                                                                Start_Lat <= 35.213  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lat > 35.213  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                        Humidity(%) > 82.500  (1.1%)
                                                            Wind_S <= 0.500  (0.9%)
                                                                Duration <= 0.541  (0.7%)
                                                                    Hour <= 15.500  (0.4%)
                                                                        Hour <= 10.000  (0.2%)
                                                                            Weather_Rain <= 0.500  (0.2%)
                                                                                Temperature(F) <= 65.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                Temperature(F) > 65.500  (0.0%)
                                                                                    Pressure(in) <= 29.035  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                    Pressure(in) > 29.035  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                            Weather_Rain > 0.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Hour > 10.000  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Hour > 15.500  (0.3%)
                                                                        Hour <= 16.500  (0.1%)
                                                                            Wind_W <= 0.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                            Wind_W > 0.500  (0.0%)
                                                                                Duration <= 0.375  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                Duration > 0.375  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Hour > 16.500  (0.2%)
                                                                            Start_Lat <= 35.262  (0.2%)
                                                                                Hour <= 18.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                Hour > 18.500  (0.0%)
                                                                                    Start_Lng <= -80.825  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                    Start_Lng > -80.825  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                            Start_Lat > 35.262  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Duration > 0.541  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Wind_S > 0.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Wind_Speed(mph) > 14.500  (0.3%)
                                                    Duration <= 0.541  (0.2%)
                                                        Wind_W <= 0.500  (0.2%)
                                                            Wind_Speed(mph) <= 16.500  (0.2%)
                                                                Wind_Speed(mph) <= 15.500  (0.0%)
                                                                    Start_Lat <= 35.275  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Start_Lat > 35.275  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                Wind_Speed(mph) > 15.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Wind_Speed(mph) > 16.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Wind_W > 0.500  (0.0%)
                                                            Start_Lng <= -80.842  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                            Start_Lng > -80.842  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Duration > 0.541  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Start_Lat > 35.278  (2.0%)
                                                Weather_Cloudy <= 0.500  (1.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Weather_Cloudy > 0.500  (0.8%)
                                                    Hour <= 17.500  (0.7%)
                                                        Start_Lat <= 35.340  (0.5%)
                                                            Pressure(in) <= 29.550  (0.5%)
                                                                Wind_W <= 0.500  (0.5%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Wind_W > 0.500  (0.0%)
                                                                    Pressure(in) <= 28.990  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                    Pressure(in) > 28.990  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Pressure(in) > 29.550  (0.0%)
                                                                Duration <= 0.458  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Duration > 0.458  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Start_Lat > 35.340  (0.1%)
                                                            Temperature(F) <= 64.500  (0.1%)
                                                                Start_Lng <= -80.740  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lng > -80.740  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                            Temperature(F) > 64.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Hour > 17.500  (0.1%)
                                                        Visibility(mi) <= 8.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Visibility(mi) > 8.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                        Junction > 0.500  (1.3%)
                                            Temperature(F) <= 55.500  (0.3%)
                                                Pressure(in) <= 29.125  (0.1%)
                                                    Wind_Speed(mph) <= 11.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Wind_Speed(mph) > 11.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                Pressure(in) > 29.125  (0.2%)
                                                    Humidity(%) <= 51.000  (0.1%)
                                                        Humidity(%) <= 43.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                        Humidity(%) > 43.500  (0.1%)
                                                            Week <= 1.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Week > 1.500  (0.0%)
                                                                Hour <= 8.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Hour > 8.500  (0.0%)  ---> [0.500, 0.500, 0.000, 0.000]
                                                    Humidity(%) > 51.000  (0.2%)
                                                        Humidity(%) <= 89.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Humidity(%) > 89.500  (0.0%)
                                                            Pressure(in) <= 29.270  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Pressure(in) > 29.270  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Temperature(F) > 55.500  (1.0%)
                                                Wind_N <= 0.500  (0.6%)
                                                    Pressure(in) <= 29.490  (0.5%)
                                                        Hour <= 6.500  (0.1%)
                                                            Start_Lat <= 35.273  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Start_Lat > 35.273  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Hour > 6.500  (0.5%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Pressure(in) > 29.490  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Wind_N > 0.500  (0.4%)
                                                    Visibility(mi) <= 5.000  (0.2%)
                                                        Month <= 4.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Month > 4.000  (0.1%)
                                                            Duration <= 0.604  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Duration > 0.604  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Visibility(mi) > 5.000  (0.3%)
                                                        Start_Lng <= -80.965  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Start_Lng > -80.965  (0.2%)
                                                            Start_Lat <= 35.351  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Start_Lat > 35.351  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Pressure(in) > 30.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                    Temperature(F) > 68.500  (7.3%)
                        Start_Lng <= -80.985  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Start_Lng > -80.985  (7.2%)
                            Start_Lng <= -80.809  (5.4%)
                                Duration <= 0.541  (4.4%)
                                    Pressure(in) <= 29.355  (3.9%)
                                        Start_Lat <= 35.123  (0.4%)
                                            Duration <= 0.487  (0.2%)
                                                Month <= 5.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Month > 5.500  (0.2%)
                                                    Wind_Speed(mph) <= 7.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Wind_Speed(mph) > 7.000  (0.1%)
                                                        Hour <= 14.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                        Hour > 14.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Duration > 0.487  (0.2%)
                                                Temperature(F) <= 89.500  (0.2%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Temperature(F) > 89.500  (0.0%)  ---> [0.000, 0.000, 0.500, 0.500]
                                        Start_Lat > 35.123  (3.5%)
                                            Pressure(in) <= 29.085  (0.3%)
                                                Start_Lng <= -80.966  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Start_Lng > -80.966  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Pressure(in) > 29.085  (3.2%)
                                                Start_Lat <= 35.229  (1.2%)
                                                    Wind_N <= 0.500  (1.0%)
                                                        Pressure(in) <= 29.130  (0.2%)
                                                            Week <= 2.500  (0.1%)
                                                                Start_Lng <= -80.850  (0.1%)
                                                                    Wind_C <= 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Wind_C > 0.500  (0.0%)
                                                                        Start_Lng <= -80.876  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Start_Lng > -80.876  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lng > -80.850  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                            Week > 2.500  (0.1%)
                                                                Duration <= 0.312  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Duration > 0.312  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Pressure(in) > 29.130  (0.8%)
                                                            Hour <= 19.500  (0.8%)
                                                                Pressure(in) <= 29.265  (0.5%)
                                                                    Start_Lng <= -80.819  (0.4%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Start_Lng > -80.819  (0.1%)
                                                                        Temperature(F) <= 78.000  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Temperature(F) > 78.000  (0.0%)
                                                                            Pressure(in) <= 29.155  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                            Pressure(in) > 29.155  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Pressure(in) > 29.265  (0.2%)
                                                                    Temperature(F) <= 79.000  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Temperature(F) > 79.000  (0.1%)
                                                                        Duration <= 0.375  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Duration > 0.375  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Hour > 19.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Wind_N > 0.500  (0.2%)
                                                        Hour <= 15.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Hour > 15.500  (0.1%)
                                                            Junction <= 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Junction > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Start_Lat > 35.229  (2.0%)
                                                    Humidity(%) <= 47.500  (0.5%)
                                                        Wind_Speed(mph) <= 16.500  (0.5%)
                                                            Duration <= 0.361  (0.2%)
                                                                Start_Lat <= 35.231  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                Start_Lat > 35.231  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Duration > 0.361  (0.3%)
                                                                Humidity(%) <= 41.500  (0.2%)
                                                                    Pressure(in) <= 29.150  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Pressure(in) > 29.150  (0.1%)
                                                                        Hour <= 14.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Hour > 14.500  (0.1%)
                                                                            Start_Lat <= 35.250  (0.0%)
                                                                                Duration <= 0.495  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                Duration > 0.495  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                            Start_Lat > 35.250  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Humidity(%) > 41.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Wind_Speed(mph) > 16.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Humidity(%) > 47.500  (1.5%)
                                                        Temperature(F) <= 69.500  (0.2%)
                                                            Hour <= 18.500  (0.1%)
                                                                Start_Lng <= -80.855  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lng > -80.855  (0.0%)
                                                                    Wind_Speed(mph) <= 16.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Wind_Speed(mph) > 16.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Hour > 18.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                        Temperature(F) > 69.500  (1.3%)
                                                            Humidity(%) <= 86.000  (1.1%)
                                                                Start_Lat <= 35.333  (1.0%)
                                                                    Duration <= 0.415  (0.6%)
                                                                        Start_Lng <= -80.842  (0.5%)
                                                                            Temperature(F) <= 73.500  (0.1%)
                                                                                Visibility(mi) <= 9.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                Visibility(mi) > 9.500  (0.0%)
                                                                                    Week <= 3.000  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                                    Week > 3.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                            Temperature(F) > 73.500  (0.4%)
                                                                                Humidity(%) <= 62.500  (0.3%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                                Humidity(%) > 62.500  (0.1%)
                                                                                    Temperature(F) <= 80.000  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                                    Temperature(F) > 80.000  (0.0%)
                                                                                        Start_Lng <= -80.854  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                        Start_Lng > -80.854  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                        Start_Lng > -80.842  (0.1%)
                                                                            Start_Lng <= -80.835  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                            Start_Lng > -80.835  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                    Duration > 0.415  (0.4%)
                                                                        Wind_E <= 0.500  (0.3%)
                                                                            Wind_C <= 0.500  (0.3%)
                                                                                Pressure(in) <= 29.125  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                                Pressure(in) > 29.125  (0.2%)
                                                                                    Humidity(%) <= 50.500  (0.0%)
                                                                                        Week <= 3.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                                        Week > 3.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                    Humidity(%) > 50.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                            Wind_C > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Wind_E > 0.500  (0.1%)
                                                                            Wind_Speed(mph) <= 7.000  (0.0%)
                                                                                Hour <= 19.000  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                                Hour > 19.000  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                                            Wind_Speed(mph) > 7.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Start_Lat > 35.333  (0.1%)
                                                                    Month <= 4.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                    Month > 4.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Humidity(%) > 86.000  (0.2%)
                                                                Week <= 1.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                                Week > 1.500  (0.2%)
                                                                    Visibility(mi) <= 9.000  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                    Visibility(mi) > 9.000  (0.0%)
                                                                        Humidity(%) <= 88.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Humidity(%) > 88.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Pressure(in) > 29.355  (0.5%)
                                        Start_Lng <= -80.819  (0.5%)
                                            Pressure(in) <= 29.530  (0.5%)
                                                Humidity(%) <= 25.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Humidity(%) > 25.500  (0.5%)
                                                    Temperature(F) <= 69.500  (0.0%)
                                                        Start_Lng <= -80.870  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lng > -80.870  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                    Temperature(F) > 69.500  (0.4%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Pressure(in) > 29.530  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                        Start_Lng > -80.819  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Duration > 0.541  (1.0%)
                                    Wind_E <= 0.500  (0.9%)
                                        Start_Lat <= 35.356  (0.9%)
                                            Weather_Thunderstorm <= 0.500  (0.9%)
                                                Humidity(%) <= 21.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Humidity(%) > 21.000  (0.9%)
                                                    Temperature(F) <= 88.500  (0.8%)
                                                        Duration <= 0.604  (0.3%)
                                                            Hour <= 13.500  (0.0%)
                                                                Visibility(mi) <= 8.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Visibility(mi) > 8.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Hour > 13.500  (0.3%)
                                                                Hour <= 18.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Hour > 18.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Duration > 0.604  (0.5%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Temperature(F) > 88.500  (0.1%)
                                                        Week <= 2.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Week > 2.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Weather_Thunderstorm > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                        Start_Lat > 35.356  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                    Wind_E > 0.500  (0.1%)
                                        Hour <= 16.000  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Hour > 16.000  (0.0%)
                                            Start_Lat <= 35.198  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Start_Lat > 35.198  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                            Start_Lng > -80.809  (1.7%)
                                Visibility(mi) <= 1.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Visibility(mi) > 1.500  (1.7%)
                                    Start_Lat <= 35.336  (1.5%)
                                        Temperature(F) <= 83.500  (1.3%)
                                            Visibility(mi) <= 2.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Visibility(mi) > 2.500  (1.3%)
                                                Wind_S <= 0.500  (0.8%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Wind_S > 0.500  (0.5%)
                                                    Give_Way <= 0.500  (0.5%)
                                                        Hour <= 16.500  (0.3%)
                                                            Start_Lng <= -80.787  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Start_Lng > -80.787  (0.1%)
                                                                Hour <= 14.000  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Hour > 14.000  (0.1%)
                                                                    Week <= 1.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Week > 1.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Hour > 16.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Give_Way > 0.500  (0.0%)
                                                        Weather_Cloudy <= 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                        Weather_Cloudy > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Temperature(F) > 83.500  (0.2%)
                                            Start_Lng <= -80.781  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Start_Lng > -80.781  (0.1%)
                                                Wind_Speed(mph) <= 11.000  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Wind_Speed(mph) > 11.000  (0.0%)
                                                    Start_Lng <= -80.759  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Start_Lng > -80.759  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Start_Lat > 35.336  (0.2%)
                                        Wind_W <= 0.500  (0.2%)
                                            Pressure(in) <= 29.290  (0.1%)
                                                Month <= 6.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Month > 6.500  (0.0%)
                                                    Temperature(F) <= 76.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Temperature(F) > 76.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Pressure(in) > 29.290  (0.1%)
                                                Wind_Speed(mph) <= 6.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Wind_Speed(mph) > 6.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                        Wind_W > 0.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
            Traffic_Signal > 0.500  (24.0%)
                Hour <= 12.500  (7.2%)
                    Station <= 0.500  (7.2%)
                        Amenity <= 0.500  (7.1%)
                            Start_Lng <= -80.850  (2.5%)
                                Hour <= 4.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                Hour > 4.500  (2.5%)
                                    Start_Lat <= 35.246  (1.6%)
                                        Humidity(%) <= 43.500  (0.2%)
                                            Week <= 0.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                            Week > 0.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Humidity(%) > 43.500  (1.4%)
                                            Start_Lng <= -80.850  (1.4%)
                                                Month <= 7.000  (1.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Month > 7.000  (0.1%)
                                                    Hour <= 10.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Hour > 10.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Start_Lng > -80.850  (0.0%)
                                                Month <= 5.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Month > 5.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Start_Lat > 35.246  (0.9%)
                                        Start_Lat <= 35.249  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                        Start_Lat > 35.249  (0.8%)
                                            Start_Lat <= 35.303  (0.3%)
                                                Hour <= 7.500  (0.1%)
                                                    Wind_N <= 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Wind_N > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Hour > 7.500  (0.3%)
                                                    Wind_Speed(mph) <= 12.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Wind_Speed(mph) > 12.500  (0.1%)
                                                        Start_Lng <= -80.883  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lng > -80.883  (0.0%)
                                                            Wind_N <= 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Wind_N > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Start_Lat > 35.303  (0.5%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Start_Lng > -80.850  (4.6%)
                                Wind_Speed(mph) <= 17.500  (4.5%)
                                    Temperature(F) <= 83.500  (4.3%)
                                        Pressure(in) <= 28.925  (0.2%)
                                            Weather_Fair <= 0.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Weather_Fair > 0.500  (0.0%)
                                                Hour <= 7.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Hour > 7.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Pressure(in) > 28.925  (4.1%)
                                            Humidity(%) <= 42.500  (0.3%)
                                                Wind_S <= 0.500  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Wind_S > 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Humidity(%) > 42.500  (3.7%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Temperature(F) > 83.500  (0.2%)
                                        Humidity(%) <= 65.000  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Humidity(%) > 65.000  (0.0%)
                                            Wind_Speed(mph) <= 6.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Wind_Speed(mph) > 6.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Wind_Speed(mph) > 17.500  (0.1%)
                                    Hour <= 7.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Hour > 7.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                        Amenity > 0.500  (0.1%)
                            Pressure(in) <= 29.315  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Pressure(in) > 29.315  (0.0%)
                                Wind_Speed(mph) <= 9.000  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                Wind_Speed(mph) > 9.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                    Station > 0.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                Hour > 12.500  (16.8%)
                    Duration <= 0.471  (5.8%)
                        Pressure(in) <= 28.765  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Pressure(in) > 28.765  (5.8%)
                            Humidity(%) <= 82.000  (5.2%)
                                Temperature(F) <= 88.500  (4.9%)
                                    Start_Lat <= 35.333  (4.5%)
                                        Sunrise_Sunset <= 0.500  (4.5%)
                                            Month <= 5.500  (3.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Month > 5.500  (1.2%)
                                                Start_Lat <= 35.197  (0.4%)
                                                    Pressure(in) <= 29.245  (0.2%)
                                                        Week <= 2.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Week > 2.500  (0.1%)
                                                            Pressure(in) <= 29.150  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Pressure(in) > 29.150  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Pressure(in) > 29.245  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Start_Lat > 35.197  (0.8%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Sunrise_Sunset > 0.500  (0.1%)
                                            Pressure(in) <= 29.260  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Pressure(in) > 29.260  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Start_Lat > 35.333  (0.4%)
                                        Pressure(in) <= 29.295  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Pressure(in) > 29.295  (0.1%)
                                            Wind_W <= 0.500  (0.1%)
                                                Start_Lat <= 35.341  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Start_Lat > 35.341  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Wind_W > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Temperature(F) > 88.500  (0.2%)
                                    Start_Lat <= 35.213  (0.1%)
                                        Start_Lat <= 35.196  (0.1%)
                                            Side <= 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Side > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Start_Lat > 35.196  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Start_Lat > 35.213  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Humidity(%) > 82.000  (0.6%)
                                Pressure(in) <= 29.395  (0.6%)
                                    Weather_Thunderstorm <= 0.500  (0.6%)
                                        Wind_Speed(mph) <= 13.500  (0.4%)
                                            Pressure(in) <= 29.090  (0.1%)
                                                Start_Lat <= 35.226  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Start_Lat > 35.226  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Pressure(in) > 29.090  (0.4%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Wind_Speed(mph) > 13.500  (0.2%)
                                            Start_Lng <= -80.890  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Start_Lng > -80.890  (0.1%)
                                                Weather_Rain <= 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Weather_Rain > 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Weather_Thunderstorm > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Pressure(in) > 29.395  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                    Duration > 0.471  (11.0%)
                        Start_Lng <= -80.955  (0.7%)
                            Humidity(%) <= 45.500  (0.3%)
                                Week <= 2.500  (0.2%)
                                    Pressure(in) <= 29.410  (0.1%)
                                        Humidity(%) <= 43.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Humidity(%) > 43.500  (0.0%)
                                            Start_Lat <= 35.290  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Start_Lat > 35.290  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Pressure(in) > 29.410  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Week > 2.500  (0.1%)
                                    Humidity(%) <= 24.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Humidity(%) > 24.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                            Humidity(%) > 45.500  (0.4%)
                                Wind_Speed(mph) <= 5.500  (0.1%)
                                    Hour <= 20.500  (0.1%)
                                        Month <= 3.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Month > 3.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Hour > 20.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                Wind_Speed(mph) > 5.500  (0.3%)
                                    Wind_Speed(mph) <= 12.500  (0.3%)
                                        Temperature(F) <= 87.500  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Temperature(F) > 87.500  (0.0%)
                                            Wind_W <= 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Wind_W > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Wind_Speed(mph) > 12.500  (0.0%)
                                        Temperature(F) <= 74.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Temperature(F) > 74.000  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                        Start_Lng > -80.955  (10.2%)
                            Start_Lat <= 35.046  (0.2%)
                                Temperature(F) <= 68.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Temperature(F) > 68.000  (0.1%)
                                    Hour <= 15.000  (0.1%)
                                        Wind_V <= 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Wind_V > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Hour > 15.000  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                            Start_Lat > 35.046  (10.1%)
                                Start_Lat <= 35.266  (6.3%)
                                    Start_Lat <= 35.222  (4.3%)
                                        Wind_Speed(mph) <= 7.500  (1.8%)
                                            Weather_Rain <= 0.500  (1.6%)
                                                Temperature(F) <= 69.500  (0.6%)
                                                    Pressure(in) <= 29.015  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Pressure(in) > 29.015  (0.6%)
                                                        Hour <= 22.000  (0.6%)
                                                            Humidity(%) <= 29.000  (0.0%)
                                                                Hour <= 16.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Hour > 16.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Humidity(%) > 29.000  (0.6%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Hour > 22.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Temperature(F) > 69.500  (0.9%)
                                                    Temperature(F) <= 70.500  (0.1%)
                                                        Hour <= 16.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Hour > 16.500  (0.1%)
                                                            Crossing <= 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Crossing > 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                    Temperature(F) > 70.500  (0.9%)
                                                        Start_Lng <= -80.878  (0.1%)
                                                            Humidity(%) <= 68.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Humidity(%) > 68.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lng > -80.878  (0.8%)
                                                            Humidity(%) <= 47.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Humidity(%) > 47.500  (0.5%)
                                                                Hour <= 13.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Hour > 13.500  (0.5%)
                                                                    Humidity(%) <= 51.500  (0.0%)
                                                                        Weather_Cloudy <= 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Weather_Cloudy > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Humidity(%) > 51.500  (0.5%)
                                                                        Wind_Speed(mph) <= 5.500  (0.2%)
                                                                            Start_Lat <= 35.091  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                            Start_Lat > 35.091  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Wind_Speed(mph) > 5.500  (0.2%)
                                                                            Pressure(in) <= 29.245  (0.1%)
                                                                                Start_Lng <= -80.773  (0.1%)
                                                                                    Wind_S <= 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                                    Wind_S > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                Start_Lng > -80.773  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                            Pressure(in) > 29.245  (0.1%)
                                                                                Side <= 0.500  (0.0%)
                                                                                    Wind_E <= 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                                    Wind_E > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                                Side > 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Weather_Rain > 0.500  (0.2%)
                                                Wind_Speed(mph) <= 5.500  (0.1%)
                                                    Temperature(F) <= 72.000  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Temperature(F) > 72.000  (0.0%)
                                                        Hour <= 17.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Hour > 17.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Wind_Speed(mph) > 5.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Wind_Speed(mph) > 7.500  (2.6%)
                                            Temperature(F) <= 67.500  (0.9%)
                                                Pressure(in) <= 29.435  (0.9%)
                                                    Weather_Cloudy <= 0.500  (0.6%)
                                                        Duration <= 0.708  (0.6%)
                                                            Start_Lat <= 35.216  (0.5%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Start_Lat > 35.216  (0.0%)
                                                                Wind_W <= 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Wind_W > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Duration > 0.708  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Weather_Cloudy > 0.500  (0.3%)
                                                        Week <= 2.500  (0.1%)
                                                            Hour <= 13.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Hour > 13.500  (0.1%)
                                                                Start_Lat <= 35.191  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lat > 35.191  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Week > 2.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Pressure(in) > 29.435  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Temperature(F) > 67.500  (1.7%)
                                                Humidity(%) <= 29.500  (0.0%)
                                                    Hour <= 15.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Hour > 15.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Humidity(%) > 29.500  (1.6%)
                                                    Temperature(F) <= 81.500  (1.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Temperature(F) > 81.500  (0.5%)
                                                        Humidity(%) <= 50.000  (0.2%)
                                                            Start_Lat <= 35.216  (0.2%)
                                                                Pressure(in) <= 29.100  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Pressure(in) > 29.100  (0.2%)
                                                                    Start_Lng <= -80.871  (0.0%)
                                                                        Start_Lng <= -80.901  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Start_Lng > -80.901  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                    Start_Lng > -80.871  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Start_Lat > 35.216  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Humidity(%) > 50.000  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Start_Lat > 35.222  (2.0%)
                                        Pressure(in) <= 28.965  (0.1%)
                                            Week <= 3.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Week > 3.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Pressure(in) > 28.965  (1.9%)
                                            Wind_Speed(mph) <= 19.500  (1.9%)
                                                Week <= 2.500  (1.0%)
                                                    Hour <= 20.500  (1.0%)
                                                        Month <= 5.500  (0.6%)
                                                            Temperature(F) <= 69.500  (0.4%)
                                                                Duration <= 0.708  (0.3%)
                                                                    Humidity(%) <= 97.000  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Humidity(%) > 97.000  (0.0%)
                                                                        Start_Lat <= 35.248  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Start_Lat > 35.248  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Duration > 0.708  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Temperature(F) > 69.500  (0.2%)
                                                                Start_Lat <= 35.237  (0.1%)
                                                                    Humidity(%) <= 67.500  (0.1%)
                                                                        Duration <= 0.666  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Duration > 0.666  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Humidity(%) > 67.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lat > 35.237  (0.1%)
                                                                    Pressure(in) <= 29.355  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Pressure(in) > 29.355  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Month > 5.500  (0.4%)
                                                            Wind_Speed(mph) <= 15.500  (0.4%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Wind_Speed(mph) > 15.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                    Hour > 20.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                Week > 2.500  (0.8%)
                                                    Temperature(F) <= 85.500  (0.7%)
                                                        Hour <= 16.500  (0.3%)
                                                            Humidity(%) <= 75.000  (0.2%)
                                                                Weather_Cloudy <= 0.500  (0.1%)
                                                                    Pressure(in) <= 29.325  (0.0%)
                                                                        Start_Lng <= -80.805  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                                        Start_Lng > -80.805  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Pressure(in) > 29.325  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Weather_Cloudy > 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Humidity(%) > 75.000  (0.1%)
                                                                Hour <= 14.500  (0.0%)
                                                                    Temperature(F) <= 68.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Temperature(F) > 68.000  (0.0%)
                                                                        Start_Lng <= -80.832  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Start_Lng > -80.832  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Hour > 14.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Hour > 16.500  (0.4%)
                                                            Humidity(%) <= 59.000  (0.2%)
                                                                Start_Lng <= -80.872  (0.0%)
                                                                    Start_Lat <= 35.235  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                    Start_Lat > 35.235  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                Start_Lng > -80.872  (0.1%)
                                                                    Temperature(F) <= 62.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                    Temperature(F) > 62.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Humidity(%) > 59.000  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Temperature(F) > 85.500  (0.1%)
                                                        Start_Lng <= -80.868  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lng > -80.868  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Wind_Speed(mph) > 19.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Start_Lat > 35.266  (3.8%)
                                    Duration <= 0.477  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                    Duration > 0.477  (3.7%)
                                        Wind_Speed(mph) <= 23.000  (3.7%)
                                            Hour <= 18.500  (2.8%)
                                                Humidity(%) <= 19.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Humidity(%) > 19.500  (2.8%)
                                                    Humidity(%) <= 63.000  (2.0%)
                                                        Pressure(in) <= 28.970  (0.1%)
                                                            Pressure(in) <= 28.950  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Pressure(in) > 28.950  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Pressure(in) > 28.970  (1.9%)
                                                            Wind_Speed(mph) <= 1.500  (0.2%)
                                                                Pressure(in) <= 29.195  (0.0%)
                                                                    Pressure(in) <= 29.165  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Pressure(in) > 29.165  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Pressure(in) > 29.195  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Wind_Speed(mph) > 1.500  (1.7%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Humidity(%) > 63.000  (0.8%)
                                                        Start_Lng <= -80.776  (0.4%)
                                                            Start_Lng <= -80.825  (0.2%)
                                                                Duration <= 0.562  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Duration > 0.562  (0.0%)
                                                                    Week <= 1.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Week > 1.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Start_Lng > -80.825  (0.2%)
                                                                Start_Lat <= 35.310  (0.1%)
                                                                    Wind_S <= 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Wind_S > 0.500  (0.0%)
                                                                        Humidity(%) <= 71.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Humidity(%) > 71.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lat > 35.310  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Start_Lng > -80.776  (0.4%)
                                                            Start_Lng <= -80.714  (0.4%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Start_Lng > -80.714  (0.0%)
                                                                Wind_C <= 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Wind_C > 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Hour > 18.500  (0.9%)
                                                Start_Lat <= 35.324  (0.7%)
                                                    Start_Lat <= 35.290  (0.1%)
                                                        Start_Lat <= 35.280  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lat > 35.280  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                    Start_Lat > 35.290  (0.6%)
                                                        Temperature(F) <= 79.000  (0.5%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Temperature(F) > 79.000  (0.1%)
                                                            Week <= 3.500  (0.0%)
                                                                Start_Lng <= -80.800  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lng > -80.800  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                            Week > 3.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Start_Lat > 35.324  (0.2%)
                                                    Start_Lng <= -80.812  (0.1%)
                                                        Start_Lng <= -80.856  (0.0%)
                                                            Wind_W <= 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Wind_W > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lng > -80.856  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Start_Lng > -80.812  (0.1%)
                                                        Pressure(in) <= 29.330  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                        Pressure(in) > 29.330  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Wind_Speed(mph) > 23.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
        Week > 4.500  (9.1%)
            Humidity(%) <= 27.500  (1.0%)
                Humidity(%) <= 25.500  (0.7%)
                    Start_Lng <= -80.796  (0.5%)
                        Start_Lat <= 35.341  (0.4%)
                            Hour <= 16.500  (0.3%)
                                Start_Lat <= 35.269  (0.1%)
                                    Pressure(in) <= 29.190  (0.0%)
                                        Start_Lat <= 35.127  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Start_Lat > 35.127  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Pressure(in) > 29.190  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                Start_Lat > 35.269  (0.2%)  ---> [1.000, 0.000, 0.000, 0.000]
                            Hour > 16.500  (0.2%)
                                Wind_S <= 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Wind_S > 0.500  (0.1%)
                                    Weather_Cloudy <= 0.500  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                    Weather_Cloudy > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Start_Lat > 35.341  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                    Start_Lng > -80.796  (0.2%)
                        Humidity(%) <= 23.500  (0.1%)
                            Hour <= 17.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                            Hour > 17.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                        Humidity(%) > 23.500  (0.2%)
                            Weather_Cloudy <= 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Weather_Cloudy > 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                Humidity(%) > 25.500  (0.3%)
                    Hour <= 13.500  (0.1%)
                        Start_Lat <= 35.221  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                        Start_Lat > 35.221  (0.0%)
                            Crossing <= 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Crossing > 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                    Hour > 13.500  (0.2%)
                        Hour <= 16.500  (0.2%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Hour > 16.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
            Humidity(%) > 27.500  (8.1%)
                Month <= 4.500  (3.2%)
                    Wind_V <= 0.500  (3.1%)
                        Start_Lng <= -80.736  (2.6%)
                            Start_Lng <= -82.940  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                            Start_Lng > -82.940  (2.6%)
                                Wind_Speed(mph) <= 24.500  (2.5%)
                                    Hour <= 2.500  (0.0%)
                                        Month <= 3.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Month > 3.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Hour > 2.500  (2.5%)
                                        Start_Lng <= -80.754  (2.1%)
                                            Pressure(in) <= 29.095  (0.4%)
                                                Start_Lat <= 35.281  (0.3%)
                                                    Temperature(F) <= 59.500  (0.1%)
                                                        Start_Lat <= 35.210  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Start_Lat > 35.210  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Temperature(F) > 59.500  (0.3%)
                                                        Start_Lat <= 35.250  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lat > 35.250  (0.0%)
                                                            Humidity(%) <= 63.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Humidity(%) > 63.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                Start_Lat > 35.281  (0.0%)
                                                    Hour <= 14.000  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Hour > 14.000  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Pressure(in) > 29.095  (1.7%)
                                                Hour <= 19.500  (1.6%)
                                                    Wind_W <= 0.500  (1.4%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Wind_W > 0.500  (0.2%)
                                                        Start_Lat <= 35.185  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Start_Lat > 35.185  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Hour > 19.500  (0.1%)
                                                    Wind_Speed(mph) <= 15.000  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Wind_Speed(mph) > 15.000  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Start_Lng > -80.754  (0.5%)
                                            Start_Lng <= -80.752  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Start_Lng > -80.752  (0.4%)
                                                Duration <= 0.667  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Duration > 0.667  (0.1%)
                                                    Duration <= 0.791  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Duration > 0.791  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Wind_Speed(mph) > 24.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Start_Lng > -80.736  (0.5%)
                            Wind_Speed(mph) <= 1.500  (0.1%)
                                Humidity(%) <= 64.000  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Humidity(%) > 64.000  (0.0%)
                                    Start_Lat <= 35.333  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                    Start_Lat > 35.333  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Wind_Speed(mph) > 1.500  (0.4%)
                                Start_Lng <= -80.735  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                Start_Lng > -80.735  (0.4%)
                                    Start_Lat <= 35.335  (0.3%)
                                        Start_Lat <= 35.202  (0.1%)
                                            Pressure(in) <= 29.160  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Pressure(in) > 29.160  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Start_Lat > 35.202  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Start_Lat > 35.335  (0.1%)
                                        Pressure(in) <= 29.170  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Pressure(in) > 29.170  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                    Wind_V > 0.500  (0.1%)
                        Traffic_Signal <= 0.500  (0.1%)
                            Weather_Cloudy <= 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                            Weather_Cloudy > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Traffic_Signal > 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                Month > 4.500  (4.9%)
                    Start_Lat <= 35.365  (4.8%)
                        Hour <= 12.500  (1.3%)
                            Hour <= 3.500  (0.2%)
                                Duration <= 0.666  (0.1%)
                                    Weather_Fair <= 0.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                    Weather_Fair > 0.500  (0.1%)
                                        Wind_Speed(mph) <= 2.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Wind_Speed(mph) > 2.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Duration > 0.666  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                            Hour > 3.500  (1.1%)
                                Week <= 5.500  (0.6%)
                                    Wind_Speed(mph) <= 13.000  (0.6%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Wind_Speed(mph) > 13.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Week > 5.500  (0.5%)
                                    Humidity(%) <= 54.000  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Humidity(%) > 54.000  (0.3%)
                                        Temperature(F) <= 76.500  (0.1%)
                                            Visibility(mi) <= 9.500  (0.0%)
                                                Pressure(in) <= 29.235  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Pressure(in) > 29.235  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Visibility(mi) > 9.500  (0.1%)
                                                Start_Lng <= -80.735  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Start_Lng > -80.735  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Temperature(F) > 76.500  (0.1%)
                                            Wind_W <= 0.500  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Wind_W > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                        Hour > 12.500  (3.5%)
                            Hour <= 17.500  (2.6%)
                                Temperature(F) <= 78.500  (0.4%)
                                    Temperature(F) <= 63.500  (0.0%)
                                        Wind_Speed(mph) <= 14.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                        Wind_Speed(mph) > 14.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Temperature(F) > 63.500  (0.4%)
                                        Junction <= 0.500  (0.3%)
                                            Weather_Thunderstorm <= 0.500  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Weather_Thunderstorm > 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Junction > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Temperature(F) > 78.500  (2.2%)
                                    Pressure(in) <= 29.315  (1.6%)
                                        Wind_W <= 0.500  (1.0%)
                                            Temperature(F) <= 87.500  (0.9%)
                                                Hour <= 14.500  (0.3%)
                                                    Humidity(%) <= 49.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Humidity(%) > 49.500  (0.1%)
                                                        Start_Lat <= 35.236  (0.1%)
                                                            Start_Lat <= 35.146  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Start_Lat > 35.146  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                        Start_Lat > 35.236  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Hour > 14.500  (0.5%)
                                                    Start_Lng <= -80.884  (0.1%)
                                                        Start_Lat <= 35.313  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lat > 35.313  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Start_Lng > -80.884  (0.4%)
                                                        Start_Lat <= 35.307  (0.4%)
                                                            Wind_S <= 0.500  (0.3%)
                                                                Temperature(F) <= 83.500  (0.2%)
                                                                    Start_Lng <= -80.809  (0.1%)
                                                                        Start_Lng <= -80.832  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                        Start_Lng > -80.832  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                    Start_Lng > -80.809  (0.1%)
                                                                        Start_Lng <= -80.784  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Start_Lng > -80.784  (0.0%)
                                                                            Duration <= 0.666  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                            Duration > 0.666  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Temperature(F) > 83.500  (0.1%)
                                                                    Wind_C <= 0.500  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                    Wind_C > 0.500  (0.1%)
                                                                        Traffic_Signal <= 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                        Traffic_Signal > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Wind_S > 0.500  (0.1%)
                                                                Start_Lat <= 35.257  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lat > 35.257  (0.0%)
                                                                    Start_Lat <= 35.296  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                    Start_Lat > 35.296  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Start_Lat > 35.307  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Temperature(F) > 87.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Wind_W > 0.500  (0.5%)
                                            Start_Lat <= 35.256  (0.4%)
                                                Start_Lat <= 35.147  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Start_Lat > 35.147  (0.4%)
                                                    Pressure(in) <= 29.230  (0.3%)
                                                        Pressure(in) <= 29.190  (0.2%)
                                                            Junction <= 0.500  (0.1%)
                                                                Humidity(%) <= 45.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                Humidity(%) > 45.500  (0.1%)
                                                                    Start_Lat <= 35.213  (0.1%)
                                                                        Wind_Speed(mph) <= 8.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Wind_Speed(mph) > 8.500  (0.0%)
                                                                            Start_Lat <= 35.177  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                            Start_Lat > 35.177  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Start_Lat > 35.213  (0.0%)
                                                                        Start_Lng <= -80.909  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                        Start_Lng > -80.909  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                            Junction > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                        Pressure(in) > 29.190  (0.2%)
                                                            Crossing <= 0.500  (0.1%)
                                                                Side <= 0.500  (0.1%)
                                                                    Temperature(F) <= 83.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                    Temperature(F) > 83.500  (0.0%)
                                                                        Start_Lat <= 35.201  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                        Start_Lat > 35.201  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                                Side > 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                            Crossing > 0.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                    Pressure(in) > 29.230  (0.0%)
                                                        Traffic_Signal <= 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Traffic_Signal > 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Start_Lat > 35.256  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                                    Pressure(in) > 29.315  (0.6%)
                                        Temperature(F) <= 81.500  (0.5%)
                                            Pressure(in) <= 29.410  (0.4%)
                                                Side <= 0.500  (0.1%)
                                                    Pressure(in) <= 29.385  (0.1%)
                                                        Start_Lng <= -80.785  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                        Start_Lng > -80.785  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                    Pressure(in) > 29.385  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Side > 0.500  (0.3%)
                                                    Start_Lng <= -80.740  (0.2%)
                                                        Start_Lng <= -80.911  (0.0%)
                                                            Start_Lng <= -80.982  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                            Start_Lng > -80.982  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                        Start_Lng > -80.911  (0.2%)
                                                            Wind_Speed(mph) <= 7.500  (0.1%)
                                                                Start_Lng <= -80.862  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                                Start_Lng > -80.862  (0.1%)
                                                                    Duration <= 0.375  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                                    Duration > 0.375  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                            Wind_Speed(mph) > 7.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                    Start_Lng > -80.740  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Pressure(in) > 29.410  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Temperature(F) > 81.500  (0.1%)
                                            Side <= 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Side > 0.500  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                            Hour > 17.500  (0.9%)
                                Month <= 5.500  (0.5%)
                                    Start_Lng <= -80.802  (0.2%)
                                        Sunrise_Sunset <= 0.500  (0.0%)
                                            Wind_S <= 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Wind_S > 0.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                        Sunrise_Sunset > 0.500  (0.1%)
                                            Weather_Fair <= 0.500  (0.0%)
                                                Start_Lng <= -80.882  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                Start_Lng > -80.882  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                            Weather_Fair > 0.500  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                    Start_Lng > -80.802  (0.4%)
                                        Traffic_Signal <= 0.500  (0.1%)
                                            Start_Lng <= -80.735  (0.1%)
                                                Start_Lat <= 35.239  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                                Start_Lat > 35.239  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Start_Lng > -80.735  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Traffic_Signal > 0.500  (0.3%)
                                            Start_Lat <= 35.194  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Start_Lat > 35.194  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Month > 5.500  (0.4%)  ---> [0.000, 1.000, 0.000, 0.000]
                    Start_Lat > 35.365  (0.1%)
                        Start_Lng <= -82.826  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                        Start_Lng > -82.826  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
    Duration > 1.000  (4.9%)
        Traffic_Signal <= 0.500  (3.9%)
            Duration <= 4.455  (3.3%)
                Duration <= 1.008  (1.8%)
                    Start_Lng <= -80.827  (1.5%)
                        Weather_Haze <= 0.500  (1.4%)
                            Humidity(%) <= 38.500  (0.3%)
                                Start_Lat <= 35.248  (0.1%)
                                    Month <= 3.500  (0.1%)
                                        Start_Lng <= -80.848  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Start_Lng > -80.848  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Month > 3.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Start_Lat > 35.248  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                            Humidity(%) > 38.500  (1.2%)
                                Visibility(mi) <= 2.500  (0.1%)
                                    Month <= 4.500  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                    Month > 4.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Visibility(mi) > 2.500  (1.1%)
                                    Wind_E <= 0.500  (0.9%)  ---> [1.000, 0.000, 0.000, 0.000]
                                    Wind_E > 0.500  (0.1%)
                                        Wind_Speed(mph) <= 8.500  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Wind_Speed(mph) > 8.500  (0.0%)
                                            Pressure(in) <= 29.100  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Pressure(in) > 29.100  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                        Weather_Haze > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                    Start_Lng > -80.827  (0.3%)
                        Hour <= 15.500  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                        Hour > 15.500  (0.2%)
                            Month <= 2.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                            Month > 2.500  (0.2%)
                                Hour <= 17.500  (0.1%)
                                    Visibility(mi) <= 8.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Visibility(mi) > 8.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                Hour > 17.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                Duration > 1.008  (1.5%)
                    Duration <= 1.312  (0.3%)
                        Pressure(in) <= 29.245  (0.2%)
                            Temperature(F) <= 59.000  (0.1%)
                                Duration <= 1.151  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Duration > 1.151  (0.0%)
                                    Month <= 3.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Month > 3.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                            Temperature(F) > 59.000  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Pressure(in) > 29.245  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                    Duration > 1.312  (1.2%)
                        Start_Lng <= -80.751  (1.1%)
                            Start_Lat <= 35.198  (0.3%)
                                Week <= 3.500  (0.2%)
                                    Start_Lng <= -80.888  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                    Start_Lng > -80.888  (0.1%)
                                        Pressure(in) <= 29.145  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Pressure(in) > 29.145  (0.1%)
                                            Week <= 0.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                            Week > 0.500  (0.1%)
                                                Temperature(F) <= 79.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Temperature(F) > 79.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                Week > 3.500  (0.1%)
                                    Wind_Speed(mph) <= 6.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                    Wind_Speed(mph) > 6.500  (0.0%)
                                        Week <= 4.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                        Week > 4.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Start_Lat > 35.198  (0.7%)
                                Hour <= 22.500  (0.7%)
                                    Hour <= 4.500  (0.1%)
                                        Wind_Speed(mph) <= 5.500  (0.1%)
                                            Start_Lat <= 35.302  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                            Start_Lat > 35.302  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Wind_Speed(mph) > 5.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Hour > 4.500  (0.5%)
                                        Wind_Speed(mph) <= 13.500  (0.4%)  ---> [1.000, 0.000, 0.000, 0.000]
                                        Wind_Speed(mph) > 13.500  (0.1%)
                                            Start_Lat <= 35.288  (0.1%)
                                                Month <= 5.500  (0.1%)  ---> [1.000, 0.000, 0.000, 0.000]
                                                Month > 5.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Start_Lat > 35.288  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                                Hour > 22.500  (0.1%)
                                    Start_Lng <= -80.831  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Start_Lng > -80.831  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                        Start_Lng > -80.751  (0.1%)
                            Humidity(%) <= 96.500  (0.1%)
                                Start_Lat <= 35.118  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                                Start_Lat > 35.118  (0.1%)  ---> [0.000, 0.000, 0.000, 1.000]
                            Humidity(%) > 96.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
            Duration > 4.455  (0.7%)
                Humidity(%) <= 77.500  (0.2%)
                    Humidity(%) <= 53.500  (0.1%)
                        Duration <= 11.479  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Duration > 11.479  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                    Humidity(%) > 53.500  (0.1%)
                        Start_Lat <= 35.239  (0.0%)
                            Duration <= 8.475  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Duration > 8.475  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                        Start_Lat > 35.239  (0.1%)  ---> [0.000, 0.000, 0.000, 1.000]
                Humidity(%) > 77.500  (0.5%)
                    Temperature(F) <= 29.900  (0.0%)
                        Hour <= 14.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                        Hour > 14.500  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                    Temperature(F) > 29.900  (0.5%)
                        Pressure(in) <= 29.080  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                        Pressure(in) > 29.080  (0.4%)
                            Hour <= 5.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                            Hour > 5.500  (0.4%)  ---> [0.000, 1.000, 0.000, 0.000]
        Traffic_Signal > 0.500  (1.0%)
            Hour <= 0.500  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
            Hour > 0.500  (1.0%)
                Wind_V <= 0.500  (0.9%)
                    Start_Lng <= -80.721  (0.9%)
                        Temperature(F) <= 29.450  (0.0%)
                            Temperature(F) <= 26.950  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Temperature(F) > 26.950  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Temperature(F) > 29.450  (0.8%)
                            Hour <= 21.500  (0.8%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Hour > 21.500  (0.1%)
                                Start_Lat <= 35.312  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Start_Lat > 35.312  (0.0%)  ---> [1.000, 0.000, 0.000, 0.000]
                    Start_Lng > -80.721  (0.0%)
                        Pressure(in) <= 29.065  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                        Pressure(in) > 29.065  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                Wind_V > 0.500  (0.1%)
                    Week <= 2.500  (0.0%)
                        Week <= 1.000  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Week > 1.000  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                    Week > 2.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
Month > 8.500  (44.9%)
    Start_Lat <= 35.783  (44.7%)
        Duration <= 0.495  (0.4%)
            Visibility(mi) <= 7.500  (0.2%)
                Start_Lng <= -80.746  (0.1%)
                    Duration <= 0.488  (0.0%)
                        Pressure(in) <= 29.430  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Pressure(in) > 29.430  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                    Duration > 0.488  (0.1%)  ---> [0.000, 0.000, 1.000, 0.000]
                Start_Lng > -80.746  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
            Visibility(mi) > 7.500  (0.2%)
                Humidity(%) <= 37.000  (0.0%)
                    Duration <= 0.475  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                    Duration > 0.475  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                Humidity(%) > 37.000  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
        Duration > 0.495  (44.3%)
            Junction <= 0.500  (41.5%)
                Temperature(F) <= 21.500  (0.1%)
                    Give_Way <= 0.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                    Give_Way > 0.500  (0.0%)  ---> [0.000, 0.500, 0.000, 0.500]
                Temperature(F) > 21.500  (41.4%)
                    Humidity(%) <= 23.500  (0.4%)
                        Pressure(in) <= 29.435  (0.3%)  ---> [0.000, 1.000, 0.000, 0.000]
                        Pressure(in) > 29.435  (0.0%)  ---> [0.000, 0.500, 0.000, 0.500]
                    Humidity(%) > 23.500  (41.0%)
                        Start_Lat <= 35.366  (40.6%)
                            Duration <= 15.334  (40.1%)
                                Start_Lat <= 35.267  (32.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Start_Lat > 35.267  (8.0%)
                                    Start_Lat <= 35.267  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                    Start_Lat > 35.267  (7.9%)
                                        Humidity(%) <= 32.500  (0.3%)
                                            Hour <= 15.500  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Hour > 15.500  (0.0%)
                                                Start_Lng <= -80.833  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                                Start_Lng > -80.833  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                        Humidity(%) > 32.500  (7.7%)
                                            Duration <= 7.492  (6.9%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Duration > 7.492  (0.7%)
                                                Hour <= 2.500  (0.0%)  ---> [0.000, 0.500, 0.000, 0.500]
                                                Hour > 2.500  (0.7%)  ---> [0.000, 1.000, 0.000, 0.000]
                            Duration > 15.334  (0.5%)
                                Duration <= 15.346  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                Duration > 15.346  (0.5%)  ---> [0.000, 1.000, 0.000, 0.000]
                        Start_Lat > 35.366  (0.4%)
                            Start_Lat <= 35.368  (0.0%)
                                Duration <= 2.883  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Duration > 2.883  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                            Start_Lat > 35.368  (0.4%)  ---> [0.000, 1.000, 0.000, 0.000]
            Junction > 0.500  (2.8%)
                Start_Lng <= -80.743  (2.6%)
                    Weather_Overcast <= 0.500  (2.6%)
                        Duration <= 0.498  (0.0%)
                            Wind_N <= 0.500  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                            Wind_N > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                        Duration > 0.498  (2.6%)
                            Temperature(F) <= 33.500  (0.3%)
                                Wind_Speed(mph) <= 5.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Wind_Speed(mph) > 5.500  (0.1%)
                                    Pressure(in) <= 29.180  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Pressure(in) > 29.180  (0.1%)
                                        Temperature(F) <= 29.000  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                        Temperature(F) > 29.000  (0.1%)
                                            Start_Lat <= 35.183  (0.0%)  ---> [0.000, 0.500, 0.000, 0.500]
                                            Start_Lat > 35.183  (0.0%)  ---> [0.000, 0.500, 0.000, 0.500]
                            Temperature(F) > 33.500  (2.3%)
                                Temperature(F) <= 73.500  (2.1%)
                                    Hour <= 22.500  (1.8%)  ---> [0.000, 1.000, 0.000, 0.000]
                                    Hour > 22.500  (0.3%)
                                        Start_Lng <= -80.965  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                        Start_Lng > -80.965  (0.3%)
                                            Start_Lat <= 35.139  (0.0%)
                                                Duration <= 1.375  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                                Duration > 1.375  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                            Start_Lat > 35.139  (0.2%)  ---> [0.000, 1.000, 0.000, 0.000]
                                Temperature(F) > 73.500  (0.2%)
                                    Start_Lat <= 35.122  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]
                                    Start_Lat > 35.122  (0.1%)
                                        Week <= 1.500  (0.0%)
                                            Start_Lng <= -80.886  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                                            Start_Lng > -80.886  (0.0%)  ---> [0.000, 0.500, 0.000, 0.500]
                                        Week > 1.500  (0.1%)  ---> [0.000, 1.000, 0.000, 0.000]
                    Weather_Overcast > 0.500  (0.0%)
                        Pressure(in) <= 30.250  (0.0%)  ---> [0.000, 0.000, 1.000, 0.000]
                        Pressure(in) > 30.250  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                Start_Lng > -80.743  (0.1%)
                    Weather_Cloudy <= 0.500  (0.1%)
                        Pressure(in) <= 29.280  (0.0%)  ---> [0.000, 0.333, 0.000, 0.667]
                        Pressure(in) > 29.280  (0.0%)  ---> [0.000, 0.500, 0.000, 0.500]
                    Weather_Cloudy > 0.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
    Start_Lat > 35.783  (0.2%)
        Visibility(mi) <= 7.500  (0.1%)
            Pressure(in) <= 28.615  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
            Pressure(in) > 28.615  (0.1%)  ---> [0.000, 0.000, 0.000, 1.000]
        Visibility(mi) > 7.500  (0.1%)
            Hour <= 16.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
            Hour > 16.500  (0.1%)
                Start_Lat <= 42.552  (0.1%)
                    Hour <= 22.500  (0.0%)  ---> [0.000, 1.000, 0.000, 0.000]
                    Hour > 22.500  (0.0%)  ---> [0.000, 0.500, 0.000, 0.500]
                Start_Lat > 42.552  (0.0%)  ---> [0.000, 0.000, 0.000, 1.000]

# Get the feature importance as a dataframe  
dt_cla_importances_df = pd.DataFrame(pd.Series(dt_cla.feature_importances_, index=X_cla.columns), columns=['Importance']).sort_values('Importance', ascending=False)
# Visualize the feature importance of the trained tree
plt.figure(figsize=(15, 10))
missing_value_graph = sns.barplot(y = dt_cla_importances_df.index, x = "Importance", data=dt_cla_importances_df, orient="h")
missing_value_graph.set_title("Feature importance by Decision Tree Classification", fontsize = 20)
missing_value_graph.set_ylabel("Features")

Text(0, 0.5, 'Features')

# 1-D pdp plot
dt_cla_pdp_goals = pdp.pdp_isolate(model=dt_cla, dataset=x_test_cla, model_features=cla_feature_names, feature='Duration')
# plot it
pdp.pdp_plot(dt_cla_pdp_goals, 'Duration')
plt.show()

# 2D Partial Dependence Plots
features_to_plot = ['Start_Lng', 'Start_Lat']
dt_cla_pdp_2D = pdp.pdp_interact(model=dt_cla, dataset=x_test_cla, model_features=cla_feature_names, features=features_to_plot)
pdp.pdp_interact_plot(pdp_interact_out=dt_cla_pdp_2D, feature_names=features_to_plot, plot_type='contour')
plt.show()

5.1.2.4. Gradient Boost Tree

# Gradient Boosting Classification
gbt_cla = GradientBoostingClassifier(learning_rate=0.1, max_depth=10, min_impurity_decrease=0.1, min_samples_leaf=2, n_estimators=100, random_state=0)
gbt_cla.fit(x_train_cla, y_train_cla)
gbt_cla_predictions = gbt_cla.predict(x_test_cla)
gbt_cla_results = classification_report(y_test_cla, gbt_cla_predictions, zero_division=True, output_dict=True)

# Confusion matrix and Classification report
draw_confusion_matrix(gbt_cla)
print(classification_report(y_test_cla, gbt_cla_predictions, zero_division=True))

# balanced_accuracy
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.balanced_accuracy_score.html#sklearn.metrics.balanced_accuracy_score
gbt_cla_balanced_accuracy = balanced_accuracy_score(y_test_cla, gbt_cla_predictions)
print(f'balanced_accuracy: {gbt_cla_balanced_accuracy}')

# ROC_AUC score
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html#sklearn.metrics.roc_auc_score
gbt_cla_roc_ovo_macro = roc_auc_score(y_test_cla, gbt_cla.predict_proba(x_test_cla), multi_class='ovo', average='macro') #Insensitive to class imbalance when average == 'macro'
gbt_cla_roc_ovr_weighted = roc_auc_score(y_test_cla, gbt_cla.predict_proba(x_test_cla), multi_class='ovr', average='weighted') #Sensitive to class imbalance even when average == 'macro'
print(f"roc_ovo_macro: {gbt_cla_roc_ovo_macro}") 
print(f"roc_ovr_weighted: {gbt_cla_roc_ovr_weighted}")
# Show feature importance as a table
eli5.show_weights(gbt_cla, feature_names = cla_feature_names)

              precision    recall  f1-score   support

           1       0.73      0.57      0.64       188
           2       0.92      0.98      0.95      3608
           3       0.63      0.23      0.34       279
           4       0.43      0.26      0.33        50

    accuracy                           0.90      4125
   macro avg       0.68      0.51      0.56      4125
weighted avg       0.88      0.90      0.89      4125

balanced_accuracy: 0.510837197070729
roc_ovo_macro: 0.8330396540638936
roc_ovr_weighted: 0.9068550384211104

Weight	Feature
0.1879 ± 0.2731	Duration
0.1120 ± 0.2316	Start_Lat
0.1065 ± 0.0987	Month
0.0969 ± 0.2870	Start_Lng
0.0739 ± 0.1390	Pressure(in)
0.0688 ± 0.0693	Humidity(%)
0.0665 ± 0.0554	Hour
0.0543 ± 0.0597	Temperature(F)
0.0512 ± 0.1057	Week
0.0461 ± 0.0298	Traffic_Signal
0.0420 ± 0.0510	Wind_Speed(mph)
0.0138 ± 0.0156	Visibility(mi)
0.0097 ± 0.0100	Side
0.0089 ± 0.0088	Weather_Fair
0.0070 ± 0.0076	Wind_V
0.0066 ± 0.0082	Wind_N
0.0065 ± 0.0318	Wind_S
0.0062 ± 0.0296	Junction
0.0059 ± 0.0100	Wind_E
0.0053 ± 0.0094	Wind_W
… 25 more …

# https://scikit-learn.org/stable/auto_examples/ensemble/plot_forest_importances.html
# Calculate Standard Deviation of each feature for all the trees
gbt_cla_importances_std = np.std([tree[0].feature_importances_ for tree in gbt_cla.estimators_], axis=0)
gbt_cla_importances = pd.Series(gbt_cla.feature_importances_, index=X_cla.columns)
gbt_cla_importances_df = pd.DataFrame(gbt_cla_importances, columns=['Importance'])
gbt_cla_importances_df['Std'] = gbt_cla_importances_std
gbt_cla_importances_df.sort_values('Importance', ascending=True, inplace=True)

fig, ax = plt.subplots(figsize=(15,10))
gbt_cla_importances_df['Importance'].plot.barh(xerr=gbt_cla_importances_df['Std'], color='cornflowerblue', ax=ax)
ax.set_title("Feature importances using MDI of Gradient Boosting Classification", fontsize = 22)
ax.set_xlabel("Mean decrease in impurity")
fig.tight_layout()

5.1.2.5. Random Forest

# Random Forest Classification
rf_cla = RandomForestClassifier(random_state=0)
rf_cla.fit(x_train_cla, y_train_cla)
rf_cla_predictions = rf_cla.predict(x_test_cla)
rf_cla_results = classification_report(y_test_cla, rf_cla_predictions, zero_division=True, output_dict=True)

# Confusion matrix and Classification report
draw_confusion_matrix(rf_cla)
print(classification_report(y_test_cla, rf_cla_predictions, zero_division=True))

# balanced_accuracy
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.balanced_accuracy_score.html#sklearn.metrics.balanced_accuracy_score
rf_cla_balanced_accuracy = balanced_accuracy_score(y_test_cla, rf_cla_predictions)
print(f'balanced_accuracy: {rf_cla_balanced_accuracy}')

# ROC_AUC score
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html#sklearn.metrics.roc_auc_score
rf_cla_roc_ovo_macro = roc_auc_score(y_test_cla, rf_cla.predict_proba(x_test_cla), multi_class='ovo', average='macro') #Insensitive to class imbalance when average == 'macro'
rf_cla_roc_ovr_weighted = roc_auc_score(y_test_cla, rf_cla.predict_proba(x_test_cla), multi_class='ovr', average='weighted') #Sensitive to class imbalance even when average == 'macro'
print(f"roc_ovo_macro: {rf_cla_roc_ovo_macro}") 
print(f"roc_ovr_weighted: {rf_cla_roc_ovr_weighted}")
# Show feature importance as a table
eli5.show_weights(rf_cla, feature_names = cla_feature_names)

              precision    recall  f1-score   support

           1       0.75      0.44      0.56       188
           2       0.91      0.99      0.95      3608
           3       0.75      0.22      0.34       279
           4       0.76      0.26      0.39        50

    accuracy                           0.90      4125
   macro avg       0.79      0.48      0.56      4125
weighted avg       0.89      0.90      0.88      4125

balanced_accuracy: 0.47705234861124424
roc_ovo_macro: 0.8897577650394922
roc_ovr_weighted: 0.910892430380735

Weight	Feature
0.1109 ± 0.0373	Duration
0.1033 ± 0.0280	Start_Lat
0.0993 ± 0.0307	Start_Lng
0.0815 ± 0.0333	Month
0.0791 ± 0.0252	Pressure(in)
0.0787 ± 0.0224	Temperature(F)
0.0772 ± 0.0218	Humidity(%)
0.0721 ± 0.0215	Hour
0.0575 ± 0.0195	Wind_Speed(mph)
0.0516 ± 0.0192	Week
0.0240 ± 0.0160	Traffic_Signal
0.0202 ± 0.0109	Visibility(mi)
0.0140 ± 0.0105	Side
0.0137 ± 0.0083	Junction
0.0117 ± 0.0081	Sunrise_Sunset
0.0115 ± 0.0090	Wind_S
0.0113 ± 0.0081	Weather_Cloudy
0.0103 ± 0.0078	Wind_W
0.0103 ± 0.0089	Weather_Fair
0.0099 ± 0.0075	Wind_N
… 25 more …

# https://scikit-learn.org/stable/auto_examples/ensemble/plot_forest_importances.html
# Calculate Standard Deviation of each feature for all the trees
rf_cla_importances_std = np.std([tree.feature_importances_ for tree in rf_cla.estimators_], axis=0)
rf_cla_importances = pd.Series(rf_cla.feature_importances_, index=X_cla.columns)
rf_cla_importances_df = pd.DataFrame(rf_cla_importances, columns=['Importance'])
rf_cla_importances_df['Std'] = rf_cla_importances_std
rf_cla_importances_df.sort_values('Importance', ascending=True, inplace=True)

fig, ax = plt.subplots(figsize=(15,10))
rf_cla_importances_df['Importance'].plot.barh(xerr=rf_cla_importances_df['Std'], color='cornflowerblue', ax=ax)
ax.set_title("Feature importances using MDI of Random Forest Classification", fontsize = 22)
ax.set_xlabel("Mean decrease in impurity")
fig.tight_layout()

5.1.2.6. XGBoost

# XGB Classification
# https://xgboost.readthedocs.io/en/latest/python/python_api.html#module-xgboost.sklearn
xgb_cla = XGBClassifier(learning_rate=0.3, max_depth=20, n_estimators=100, eval_metric='mlogloss', random_state=0)
xgb_cla.fit(x_train_cla, y_train_cla)
xgb_cla_predictions = xgb_cla.predict(x_test_cla)
xgb_cla_results = classification_report(y_test_cla, xgb_cla_predictions, zero_division=True, output_dict=True)

# Confusion matrix and Classification report
draw_confusion_matrix(xgb_cla)
print(classification_report(y_test_cla, xgb_cla_predictions, zero_division=True))

# balanced_accuracy
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.balanced_accuracy_score.html#sklearn.metrics.balanced_accuracy_score
xgb_cla_balanced_accuracy = balanced_accuracy_score(y_test_cla, xgb_cla_predictions)
print(f'balanced_accuracy: {xgb_cla_balanced_accuracy}')

# ROC_AUC score
# https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html#sklearn.metrics.roc_auc_score
xgb_cla_roc_ovo_macro = roc_auc_score(y_test_cla, xgb_cla.predict_proba(x_test_cla), multi_class='ovo', average='macro') #Insensitive to class imbalance when average == 'macro'
xgb_cla_roc_ovr_weighted = roc_auc_score(y_test_cla, xgb_cla.predict_proba(x_test_cla), multi_class='ovr', average='weighted') #Sensitive to class imbalance even when average == 'macro'
print(f"roc_ovo_macro: {xgb_cla_roc_ovo_macro}") 
print(f"roc_ovr_weighted: {xgb_cla_roc_ovr_weighted}")
# Show feature importance as a table
eli5.show_weights(xgb_cla, feature_names = cla_feature_names)

              precision    recall  f1-score   support

           1       0.73      0.66      0.69       188
           2       0.93      0.97      0.95      3608
           3       0.58      0.29      0.38       279
           4       0.70      0.42      0.53        50

    accuracy                           0.91      4125
   macro avg       0.73      0.59      0.64      4125
weighted avg       0.89      0.91      0.90      4125

balanced_accuracy: 0.5864639787651699
roc_ovo_macro: 0.871496232553084
roc_ovr_weighted: 0.9148482266513861

Weight	Feature
0.0932	Month
0.0822	Duration
0.0629	Traffic_Signal
0.0513	Wind_V
0.0425	Amenity
0.0405	Stop
0.0383	Week
0.0371	Weather_Overcast
0.0364	Weather_Thunder
0.0320	Side
0.0316	Junction
0.0301	Wind_E
0.0266	Hour
0.0262	Sunrise_Sunset
0.0260	Visibility(mi)
0.0254	Railway
0.0252	Start_Lat
0.0247	Weather_Fair
0.0227	Humidity(%)
0.0213	Wind_W
… 25 more …

# Get the feature importance as a dataframe  
xgb_cla_importances_df = pd.DataFrame(pd.Series(xgb_cla.feature_importances_, index=X_cla.columns), columns=['Importance']).sort_values('Importance', ascending=False)
# Visualize the feature importance of the trained tree
plt.figure(figsize=(15, 10))
missing_value_graph = sns.barplot(y = xgb_cla_importances_df.index, x = "Importance", data=xgb_cla_importances_df, orient="h")
missing_value_graph.set_title("Feature importance by XGB Classification", fontsize = 20)
missing_value_graph.set_ylabel("Features")

Text(0, 0.5, 'Features')

# 1-D pdp plot
xgb_cla_pdp_goals = pdp.pdp_isolate(model=xgb_cla, dataset=x_test_cla, model_features=cla_feature_names, feature='Crossing')
# plot it
pdp.pdp_plot(xgb_cla_pdp_goals, 'Crossing')
plt.show()

# 2D Partial Dependence Plots
features_to_plot = ['Side', 'Duration']
xgb_cla_pdp_2D = pdp.pdp_interact(model=xgb_cla, dataset=x_test_cla, model_features=cla_feature_names, features=features_to_plot)
pdp.pdp_interact_plot(pdp_interact_out=xgb_cla_pdp_2D, feature_names=features_to_plot, plot_type='contour')
plt.show()

5.1.2.8. Model Comparison

# Gather all the classification performance in one table
cla_results = pd.DataFrame([
    (logistic_balanced_accuracy, logistic_reg_results['accuracy'], logistic_reg_results['weighted avg']['precision'], logistic_reg_results['weighted avg']['recall'], logistic_reg_results['weighted avg']['f1-score'], logistic_roc_ovo_macro, logistic_roc_ovr_weighted),
    
    (dt_cla_balanced_accuracy, dt_cla_results['accuracy'], dt_cla_results['weighted avg']['precision'], dt_cla_results['weighted avg']['recall'], dt_cla_results['weighted avg']['f1-score'], dt_cla_roc_ovo_macro, dt_cla_roc_ovr_weighted),
    (gbt_cla_balanced_accuracy, gbt_cla_results['accuracy'], gbt_cla_results['weighted avg']['precision'], gbt_cla_results['weighted avg']['recall'], gbt_cla_results['weighted avg']['f1-score'], gbt_cla_roc_ovo_macro, gbt_cla_roc_ovr_weighted),
    (rf_cla_balanced_accuracy, rf_cla_results['accuracy'], rf_cla_results['weighted avg']['precision'], rf_cla_results['weighted avg']['recall'], rf_cla_results['weighted avg']['f1-score'], rf_cla_roc_ovo_macro, rf_cla_roc_ovr_weighted),
    (xgb_cla_balanced_accuracy, xgb_cla_results['accuracy'], xgb_cla_results['weighted avg']['precision'], xgb_cla_results['weighted avg']['recall'], xgb_cla_results['weighted avg']['f1-score'], xgb_cla_roc_ovo_macro, xgb_cla_roc_ovr_weighted), 
    ], 
    columns=['Accuracy(Balanced)', 'Accuracy','Precision(Weighted_avg)', 'Recall(Weighted_avg)', 'F1-score(Weighted_avg)', 'Roc_ovo(macro)', 'Roc_ovr(weighted)'], 
    index= ['Logistics Regression',
            'Decision Tree',
            'Gradient Boosting Tree',
            'Random Forest',
            'XGBoost',
            ])

cla_results.sort_values(by=['F1-score(Weighted_avg)'], ascending=False)

	Accuracy(Balanced)	Accuracy	Precision(Weighted_avg)	Recall(Weighted_avg)	F1-score(Weighted_avg)	Roc_ovo(macro)	Roc_ovr(weighted)
XGBoost	0.586464	0.906909	0.893089	0.906909	0.895685	0.871496	0.914848
Gradient Boosting Tree	0.510837	0.901576	0.884886	0.901576	0.885421	0.833040	0.906855
Random Forest	0.477052	0.902303	0.890553	0.902303	0.881792	0.889758	0.910892
Logistics Regression	0.340166	0.900681	0.854026	0.900681	0.872866	0.794588	0.920353
Decision Tree	0.540442	0.866909	0.868968	0.866909	0.867921	0.693471	0.729585

5.1.2.9. Deal with Imbalanced data

# Form the train data for sampling
train_cla_df = pd.concat([x_train_cla, y_train_cla], axis=1)

# Over-sampling and Under-sampling 
size_l = len(train_cla_df[train_cla_df["Severity"]==2].index)
size_s = len(train_cla_df[train_cla_df["Severity"]==1].index)

train_cla_over = pd.DataFrame()
train_cla_under = pd.DataFrame()

for i in range(1,5):
    class_df = train_cla_df[train_cla_df["Severity"]==i]
    train_cla_over = train_cla_over.append(class_df.sample(size_l, random_state=1, replace=True))
    train_cla_under = train_cla_under.append(class_df.sample(size_s, random_state=1, replace=True))

print(f'Over-sampling: \n{train_cla_over.Severity.value_counts()}')
print(f'Under-sampling: \n{train_cla_under.Severity.value_counts()}')

Over-sampling: 
1    5411
2    5411
3    5411
4    5411
Name: Severity, dtype: int64
Under-sampling: 
1    282
2    282
3    282
4    282
Name: Severity, dtype: int64

# Try on over-sampling data
# XGB Classification
xgb_cla = XGBClassifier(learning_rate=0.3, max_depth=20, n_estimators=100, eval_metric='mlogloss', random_state=0)
xgb_cla.fit(train_cla_over.drop('Severity', axis=1), train_cla_over['Severity'])
xgb_cla_predictions_over = xgb_cla.predict(x_test_cla)

# Confusion matrix and Classification report
draw_confusion_matrix(xgb_cla)
print(classification_report(y_test_cla, xgb_cla_predictions_over, zero_division=True))

              precision    recall  f1-score   support

           1       0.59      0.74      0.66       188
           2       0.94      0.94      0.94      3608
           3       0.44      0.41      0.42       279
           4       0.45      0.48      0.47        50

    accuracy                           0.89      4125
   macro avg       0.61      0.64      0.62      4125
weighted avg       0.89      0.89      0.89      4125

# Try on under-sampling data
# XGB Classification
xgb_cla = XGBClassifier(learning_rate=0.3, max_depth=20, n_estimators=100, eval_metric='mlogloss', random_state=0)
xgb_cla.fit(train_cla_under.drop('Severity', axis=1), train_cla_under['Severity'])
xgb_cla_predictions_under = xgb_cla.predict(x_test_cla)

# Confusion matrix and Classification report
draw_confusion_matrix(xgb_cla)
print(classification_report(y_test_cla, xgb_cla_predictions_under, zero_division=True))

              precision    recall  f1-score   support

           1       0.34      0.77      0.48       188
           2       0.97      0.69      0.81      3608
           3       0.19      0.64      0.29       279
           4       0.14      0.58      0.23        50

    accuracy                           0.69      4125
   macro avg       0.41      0.67      0.45      4125
weighted avg       0.88      0.69      0.75      4125

It seems Over-sampling and Under-sampling technique doesn't help to improve the model performance. Need to explore more methods.

5.1.2.10. Severity Prediction Visualization

# Charlotte latitude and longitude values
# https://www.latlong.net/place/charlotte-nc-usa-1947.html
charlotte_lat = 35.2271
charlotte_long = -80.843124

# Generate a map of charlotte
charlotte_map = folium.Map(location=[charlotte_lat, charlotte_long], zoom_start=12)

# Instantiate a mark cluster object for the incidents in the dataframe
accidents = folium.plugins.MarkerCluster().add_to(charlotte_map)

# Loop through the dataframe and add each data point to the mark cluster
for lat, lng, label in zip(x_test_cla['Start_Lat'], x_test_cla['Start_Lng'], xgb_cla_predictions.astype(str)):
    if label == '4':
        folium.Marker(
            location=[lat, lng],
            icon=folium.Icon(color="red", icon="warning-sign"), #https://getbootstrap.com/docs/3.3/components/
            popup=label,
            ).add_to(accidents)
    elif label == '3':
        folium.Marker(
            location=[lat, lng],
            icon=folium.Icon(color="lightred", icon="warning-sign"),
            popup=label,
            ).add_to(accidents)
    elif label == '2':
        folium.Marker(
            location=[lat, lng],
            icon=folium.Icon(color="orange", icon="warning-sign"),
            popup=label,
            ).add_to(accidents)
    elif label == '1':
        folium.Marker(
            location=[lat, lng],
            icon=folium.Icon(color="beige", icon="warning-sign"),
            popup=label,
            ).add_to(accidents)
# Display map
charlotte_map

Make this Notebook Trusted to load map: File -> Trust Notebook

5.3. Severity Prediction

5.3.1 Parameter Tuning

# Decision Tree Classification
dtc_pipe = make_pipeline(DecisionTreeClassifier(random_state=0))
dtc_param = {
    'decisiontreeclassifier__max_depth': [5, 10, 20],
    'decisiontreeclassifier__min_samples_leaf': [2, 5, 10],
    'decisiontreeclassifier__min_impurity_decrease': [0.1, 0.2, 0.5]
}

dtc_search = GridSearchCV(dtc_pipe, 
                         dtc_param,
                         scoring="f1_weighted",
                         n_jobs=-1,
                         cv = 5)
dtc_search.fit(X_cla, Y_cla)
print(f'Best Params: {dtc_search.best_params_} \nBest score: {dtc_search.best_score_}')

Best Params: {'decisiontreeclassifier__max_depth': 5, 'decisiontreeclassifier__min_impurity_decrease': 0.1, 'decisiontreeclassifier__min_samples_leaf': 2} 
Best score: 0.8161116460040716

# Gradient Boosting Classification
gbtc_pipe = make_pipeline(GradientBoostingClassifier(random_state=0))
gbtc_param = {
    'gradientboostingclassifier__learning_rate': [0.1, 0.3, 0.7],
    'gradientboostingclassifier__max_depth': [5, 10, 20],
    'gradientboostingclassifier__n_estimators': [50, 100, 200],
    'gradientboostingclassifier__min_samples_leaf': [2, 5, 10],
    'gradientboostingclassifier__min_impurity_decrease': [0.1, 0.2, 0.5]
}

gbtc_search = GridSearchCV(gbtc_pipe, 
                         gbtc_param,
                         scoring="f1_weighted",
                         n_jobs=-1,
                         cv = 5)
gbtc_search.fit(X_cla, Y_cla)
print(f'Best Params: {gbtc_search.best_params_} \nBest score: {gbtc_search.best_score_}')

Best Params: {'gradientboostingclassifier__learning_rate': 0.3, 'gradientboostingclassifier__max_depth': 20, 'gradientboostingclassifier__min_impurity_decrease': 0.5, 'gradientboostingclassifier__min_samples_leaf': 2, 'gradientboostingclassifier__n_estimators': 50} 
Best score: 0.7733029627536956

# Random Forest Classification
rfc_pipe = make_pipeline(RandomForestClassifier(random_state=0))
rfc_param = {
    'randomforestclassifier__max_depth': [5, 10, 20],
    'randomforestclassifier__n_estimators': [50, 100, 200],
    'randomforestclassifier__min_impurity_decrease': [0.1, 0.2, 0.5],
    'randomforestclassifier__min_samples_leaf': [2, 5, 10],
}

rfc_search = GridSearchCV(rfc_pipe, 
                         rfc_param,
                         scoring="f1_weighted",
                         n_jobs=-1,
                         cv = 5)
rfc_search.fit(X_cla, Y_cla)
print(f'Best Params: {rfc_search.best_params_} \nBest score: {rfc_search.best_score_}')

Best Params: {'randomforestclassifier__max_depth': 5, 'randomforestclassifier__min_impurity_decrease': 0.1, 'randomforestclassifier__min_samples_leaf': 2, 'randomforestclassifier__n_estimators': 50} 
Best score: 0.8161116460040716

# XGB Classification
# https://xgboost.readthedocs.io/en/latest/tutorials/param_tuning.html
# https://stats.stackexchange.com/questions/243207/what-is-the-proper-usage-of-scale-pos-weight-in-xgboost-for-imbalanced-datasets
xgbc_pipe = make_pipeline(XGBClassifier(eval_metric='mlogloss', random_state=0))
xgbc_param = {
              'xgbclassifier__learning_rate': [0.1, 0.3, 0.7],
              'xgbclassifier__max_depth': [5, 10, 20],
              'xgbclassifier__n_estimators': [50, 100, 200],
              'xgbclassifier__scale_pos_weight': [5, 10, 30]}

xgbc_search = GridSearchCV(xgbc_pipe, 
                         xgbc_param,
                         scoring="f1_weighted",
                         n_jobs=-1,
                         cv = 5)
xgbc_search.fit(X_cla, Y_cla)
print(f'Best Params: {xgbc_search.best_params_} \nBest score: {xgbc_search.best_score_}')

[22:24:20] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.5.0/src/learner.cc:576: 
Parameters: { "scale_pos_weight" } might not be used.

  This could be a false alarm, with some parameters getting used by language bindings but
  then being mistakenly passed down to XGBoost core, or some parameter actually being used
  but getting flagged wrongly here. Please open an issue if you find any such cases.


Best Params: {'xgbclassifier__learning_rate': 0.1, 'xgbclassifier__max_depth': 5, 'xgbclassifier__n_estimators': 50, 'xgbclassifier__scale_pos_weight': 5} 
Best score: 0.7752979277991813

5.3.2 Model Comparison

pd.DataFrame([ dtc_search.best_score_, gbtc_search.best_score_, rfc_search.best_score_, xgbc_search.best_score_], 
             columns=['F1-score'],
             index= ['Decision Tree',
                    'Gradient Boosting Tree',
                    'Random Forest',
                    'XGBoost'
                   ]).sort_values(by=['F1-score'], ascending=False)

	F1-score
Decision Tree	0.816112
Random Forest	0.816112
XGBoost	0.775298
Gradient Boosting Tree	0.773303

Classification Results

Decision Tree

Best Params: {'decisiontreeclassifiermax_depth': 5, 'decisiontreeclassifiermin_impurity_decrease': 0.2, 'decisiontreeclassifier__min_samples_leaf': 2} Best score: 0.816112

Gradient Boosting Algorithm

Best Params: {'gradientboostingclassifierlearning_rate': 0.3, 'gradientboostingclassifiermax_depth': 20, 'gradientboostingclassifiermin_impurity_decrease': 0.5, 'gradientboostingclassifiermin_samples_leaf': 2, 'gradientboostingclassifier__n_estimators': 50} Best score: 0.773303

Random Forest Algorithm

Best Params: {'randomforestclassifiermax_depth': 5, 'randomforestclassifiermin_impurity_decrease': 0.1, 'randomforestclassifiermin_samples_leaf': 2, 'randomforestclassifiern_estimators': 50} Best score: 0.816112

XGB Classifier

Best Params:{'xgbclassifierlearning_rate': 0.1, 'xgbclassifiermax_depth': 5, 'xgbclassifiern_estimators': 50, 'xgbclassifierscale_pos_weight': 5} Best score: 0.775298

6. Conclusion

After performing Machine Learning models like, Random Forest, Decision Tree, XGB Classifier, Gradient Boosting algorithms on this dataset, we conclude that XG Boost Algorithm performs well with 90% of high positive accuracy rate.

Even after tuning the parameters of these algorithms, we got that, Decision Tree and Random Forest Model gave us with 81% positive accuracy rate comparing to other. So, finally for this dataset, XGBoost Model is the best for Car Accident Severity Prediction.

7. Model Explainability

LIME

#Modelling Part
#Splitting the data
#Lightbgm-Extreme Gradient Boosting Method
X = data_modelling_df.drop(["Severity"],axis=1)
y = data_modelling_df["Severity"]
# spliting the data to train and validation set
# train and test data
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=101,stratify=y)
# shape of the data of train and validation set
print('Shape of the X_train {}'.format(X_train.shape))
print('Shape of the y_train {}'.format(y_train.shape))
print('Shape of the X_test {}'.format(X_test.shape))
print('Shape of the y_test {}'.format(y_test.shape))
#The shape of our data on train and validation. we will be using Lightgbm Algorithm to build our model.

Shape of the X_train (8079, 37)
Shape of the y_train (8079,)
Shape of the X_test (2693, 37)
Shape of the y_test (2693,)

#Using Lightgbm Algorithm-Extreme Gradient Boosting Method
params = {}
params["objective"] = "binary"
params['metric'] = 'auc_mu'
params["max_depth"] = -1
params["num_leaves"] = 10
params["min_data_in_leaf"] = 20
params["learning_rate"] = 0.03
params["bagging_fraction"] = 0.9
params["feature_fraction"] = 0.35
params["feature_fraction_seed"] = 20
params["bagging_freq"] = 10
params["bagging_seed"] = 30
params["'min_child_weight'"] = 0.01
params["lambda_l1"] = 0.01
params["verbosity"] = -1

from lightgbm import LGBMClassifier # intializing the model
model = LGBMClassifier(**params)
# fitting the model
model.fit(X_train, y_train)

[LightGBM] [Warning] min_child_weight is set=0.001, min_child_weight=0.01 will be ignored. Current value: min_child_weight=0.001
[LightGBM] [Warning] feature_fraction is set=0.35, colsample_bytree=1.0 will be ignored. Current value: feature_fraction=0.35
[LightGBM] [Warning] min_data_in_leaf is set=20, min_child_samples=20 will be ignored. Current value: min_data_in_leaf=20
[LightGBM] [Warning] lambda_l1 is set=0.01, reg_alpha=0.0 will be ignored. Current value: lambda_l1=0.01
[LightGBM] [Warning] bagging_fraction is set=0.9, subsample=1.0 will be ignored. Current value: bagging_fraction=0.9
[LightGBM] [Warning] bagging_freq is set=10, subsample_freq=0 will be ignored. Current value: bagging_freq=10

LGBMClassifier("'min_child_weight'"=0.01, bagging_fraction=0.9, bagging_freq=10,
               bagging_seed=30, feature_fraction=0.35, feature_fraction_seed=20,
               lambda_l1=0.01, learning_rate=0.03, metric='auc_mu',
               min_data_in_leaf=20, num_leaves=10, objective='binary',
               verbosity=-1)

pred = model.predict_proba(X_train)
pred
pred = pd.DataFrame(pred[:,1])
pred.head(10)
 

	0
0	0.987100
1	0.973316
2	0.989322
3	0.986046
4	0.985822
5	0.987593
6	0.988131
7	0.994213
8	0.987010
9	0.990157

y_train = pd.DataFrame(y_train)
y_train

	Severity
1134	2
2156	2
5164	2
378	2
2551	2
...	...
4746	2
10482	2
7998	2
3027	2
5349	2

8079 rows × 1 columns

X_train.head()

	Start_Lat	Start_Lng	Side	Temperature(F)	Humidity(%)	Pressure(in)	Visibility(mi)	Wind_Speed(mph)	Amenity	Crossing	Junction	No_Exit	Railway	Roundabout	Station	Stop	Traffic_Calming	Traffic_Signal	Sunrise_Sunset	Duration	Month	Week	Hour	Weather_Fair	Weather_Cloudy	Weather_Snow	Weather_Haze	Weather_Rain	Weather_Thunderstorm	Weather_Windy	Weather_Thunder	Wind_C	Wind_E	Wind_N	Wind_S	Wind_V	Wind_W
1134	28.355466	-81.239637	1	83.0	60.0	29.89	10.0	6.0	0	1	0	0	0	0	0	0	0	0	0	9.791667	10	1	13	0	1	0	0	0	0	0	0	0	0	1	0	0	0
2156	28.368061	-81.339007	1	78.0	60.0	29.88	10.0	10.0	0	0	0	0	0	0	0	0	0	0	0	5.597500	9	2	15	0	1	0	0	0	0	0	0	0	0	1	0	0	0
5164	28.567801	-81.207809	1	77.0	45.0	30.10	10.0	12.0	0	1	0	0	0	0	0	0	0	0	0	5.934444	11	1	13	1	0	0	0	0	0	0	0	0	0	1	0	0	0
378	28.489483	-81.458757	1	59.0	93.0	29.99	10.0	5.0	0	0	0	0	0	0	0	0	0	0	1	2.275000	12	5	5	0	1	0	0	0	0	0	0	0	0	1	0	0	0
2551	28.430379	-81.404783	1	82.0	74.0	29.91	10.0	7.0	0	0	0	0	0	0	0	0	0	0	0	36.551389	9	3	11	0	1	0	0	0	0	0	0	0	0	0	1	0	0

X_test.head()

	Start_Lat	Start_Lng	Side	Temperature(F)	Humidity(%)	Pressure(in)	Visibility(mi)	Wind_Speed(mph)	Amenity	Crossing	Junction	No_Exit	Railway	Roundabout	Station	Stop	Traffic_Calming	Traffic_Signal	Sunrise_Sunset	Duration	Month	Week	Hour	Weather_Fair	Weather_Cloudy	Weather_Snow	Weather_Haze	Weather_Rain	Weather_Thunderstorm	Weather_Windy	Weather_Thunder	Wind_C	Wind_E	Wind_N	Wind_S	Wind_V	Wind_W
2334	28.482594	-81.330862	1	71.0	57.0	30.06	10.0	7.0	0	1	0	0	0	0	1	0	0	0	0	8.952500	12	5	15	0	1	0	0	0	0	0	0	0	1	0	0	0	0
9109	28.371530	-81.425670	1	84.0	61.0	29.97	10.0	7.0	0	0	0	0	0	0	0	0	0	0	0	0.500000	6	4	11	0	1	0	0	0	0	0	0	0	1	0	0	0	0
9239	28.536930	-81.375040	1	73.0	37.0	30.34	10.0	21.0	0	0	0	0	0	0	0	0	0	0	0	0.487778	3	6	13	0	1	0	0	0	0	1	0	0	1	0	0	0	0
617	28.565166	-81.207339	0	84.0	55.0	29.97	10.0	20.0	0	0	0	0	0	0	0	0	0	0	0	12.057500	9	1	13	0	1	0	0	0	0	0	0	0	1	0	0	0	0
9959	28.547080	-81.270170	1	83.0	46.0	29.86	10.0	17.0	0	0	0	0	0	0	0	0	0	0	0	0.488333	2	3	18	1	0	0	0	0	0	0	0	0	0	0	0	0	1

print(y_train.shape)
print(pred.shape)
print(y_test.shape)

(8079, 1)
(8079, 1)
(2693,)

# predicting the likelihood for the validation set
y_pred = model.predict_proba(X_test)[:, 1]
y_pred = pd.DataFrame(y_pred)
y_pred

	0
0	0.991329
1	0.742801
2	0.545205
3	0.992614
4	0.570852
...	...
2688	0.656544
2689	0.988044
2690	0.990723
2691	0.990287
2692	0.971127

2693 rows × 1 columns

#roc auc score
y_prob_pred = model.predict_proba(X_test)
ROC = roc_auc_score(y_test, y_prob_pred, multi_class='ovo', average='weighted')
# checking the roc_auc_curve
print('AUC score of the model is {}'.format(ROC))

AUC score of the model is 0.9495456584321436

import lime
from lime.lime_tabular import LimeTabularExplainer
#Prediction Class Names
class_names = [0,1]
#instantiate the explanations for the data set
k=5
#LIME Explainer
limeexplainer = lime.lime_tabular.LimeTabularExplainer(X_test.values,class_names=class_names, feature_names = X_test.columns, discretize_continuous = True)
idx=100 # the rows of the dataset
explainable_exp = limeexplainer.explain_instance(X_test.values[idx], model.predict_proba, num_features=k, labels=class_names)
explainable_exp.show_in_notebook(show_table=True)